zusammenfassung - OPUS Würzburg

Transcription

Petrogenesis of the Mesoproterozoic anorthosite, syenite
and carbonatite suites of NW Namibia and their
contribution to the metasomatic formation of the
Swartbooisdrif sodalite deposits
Dissertation zur Erlangung des
naturwissenschaftlichen Doktorgrades
der Bayerischen Julius-Maximilians-Universität Würzburg
vorgelegt von
Kirsten Drüppel
aus Würzburg
Würzburg 2003
Eingereicht am:
1. Gutachter:
2. Gutachter:
der Dissertation
1. Prüfer:
2. Prüfer:
der mündlichen Prüfung
Tag der mündlichen Prüfung:
Doktorurkunde ausgehändigt am:
Drüppel, K. (2003): Petrogenesis of the Mesoproterozoic anorthosite, syenite and carbonatite
suites of NW Namibia and their contribution to the metasomatic formation of the Swartbooisdrif
sodalite deposits. Dr. rer. nat. thesis, University of Würzburg, 345 + 199 pp., 140 Figs., 24
Tables in the text, 4 Plates.
NATUR IST MAßLOS
(FRITZ MAUTHNER, 1901)
Table of contents
1
TABLE OF CONTENTS
ZUSAMMENFASSUNG .............................................................................................................. 7
ABSTRACT ................................................................................................................................. 12
ACKNOWLEDGEMENTS........................................................................................................ 15
1 INTRODUCTION.................................................................................................................... 17
1.1 LOCATION OF THE STUDY AREA ............................................................................................ 17
1.2 GEOLOGICAL FRAMEWORK .................................................................................................. 17
1.2.1 The metamorphic Epupa Complex ...............................................................................17
1.2.2 The anorthositic Kunene Intrusive Complex................................................................18
1.2.3 The syenite-carbonatite-lamprophyre suites of Swartbooisdrif and Epembe ..............21
1.3 AVAILABLE GEOCHRONOLOGICAL DATA .............................................................................. 22
1.4 PREVIOUS WORK ON THE MAGMATIC ROCKS OF NW NAMIBIA AND SW ANGOLA ............... 24
1.5 METHODS OF INVESTIGATION ............................................................................................... 26
1.6 AIMS OF THE STUDY ............................................................................................................. 26
PART I: THE GEOLOGY OF THE SODALITE MINING AREA
AT SWARTBOOISDRIF ............................................................................................. 29
2 PHYSIOGRAPHY OF THE AREA MAPPED..................................................................... 29
2.1 LOCATION OF THE AREA MAPPED ......................................................................................... 29
2.2 GEOGRAPHIC AND GEOMORPHOLOGIC OVERVIEW ................................................................ 29
2.3 EXPLORATION HISTORY ........................................................................................................ 31
3 GEOLOGY OF THE AREA MAPPED................................................................................. 32
3.1 GEOLOGICAL OVERVIEW ...................................................................................................... 32
3.2 ROCK EXPOSURE .................................................................................................................. 34
3.3 FIELD OBSERVATIONS........................................................................................................... 35
3.3.1 The white anorthosite suite [A,w] ................................................................................ 35
3.3.2 The dark anorthosite suite............................................................................................ 36
3.3.2.1 Anorthosites and leucogabbronorites [A,m] .........................................................36
3.3.2.2 Leucotroctolites [T,m]........................................................................................... 37
3.3.2.3 Brecciated anorthosites [A,b]................................................................................38
3.3.3 Ilmenite-magnetite ore [IM]......................................................................................... 38
3.3.4 Syenites [S,a] and [S,f] ................................................................................................ 40
3.3.5 Nepheline Syenite [S,ne] ..............................................................................................42
3.3.6 Carbonatites and associated rock types.......................................................................44
3.3.6.1 Carbonatitic breccia...............................................................................................44
3.3.6.1.1 Layered carbonatitic breccia [CB,l] ............................................................... 44
3.3.6.1.2 Sodalite-rich carbonatitic breccia [CB,so] ..................................................... 46
3.3.6.1.3 Massive carbonatitic breccia [CB,m] ............................................................. 52
3.3.6.2 K-feldspar-biotite-ilmenite-calcite rock ................................................................ 54
3.3.6.3 Ferrocarbonatite .................................................................................................... 55
3.3.7 Dolerite [Dol] .............................................................................................................. 56
3.3.8 Trachyte [Tr] and Andesite [An] ................................................................................. 56
2
Table of contents
3.3.9 Sedimentary rocks ........................................................................................................ 57
3.3.9.1 Unconsolidated deposits of fluvial terraces [Qtz,p].............................................. 57
3.3.9.2 Carbonate-cemented deposits of fluvial terraces [Cgl]......................................... 58
3.3.9.3 River floor deposits [All] ...................................................................................... 58
3.3.9.4 Anorthositic talus deposits .................................................................................... 59
3.3.9.5 Mining waste piles ................................................................................................ 59
3.4 ECONOMIC POTENTIAL ......................................................................................................... 60
3.4.1 Sodalite deposits .......................................................................................................... 60
3.4.2 Leucotroctolites............................................................................................................ 61
3.5 TECTONICS ........................................................................................................................... 62
3.5.1 Tectonic overview ........................................................................................................ 62
3.5.1.1 The anorthosite problem ....................................................................................... 62
3.5.1.2 The structural setting of carbonatites .................................................................... 63
3.5.2 Structural investigations of the mapped sheet and related areas ................................ 64
3.5.2.1 The Zebra Mountain massif .................................................................................. 64
3.5.2.2 The cogenetic syenite intrusions ........................................................................... 69
3.5.2.3 The carbonatite dyke swarms................................................................................ 70
3.5.2.4 Dolerites ................................................................................................................ 74
3.5.3 Fault system ................................................................................................................. 75
PART II: PETROGENESIS AND EVOLUTION OF THE
KUNENE INTRUSIVE COMPLEX AND THE SPATIALLY
ASSOCIATED FELSIC SUITE .............................................................................. 77
4 GENERAL REMARKS .......................................................................................................... 77
5 PETROGRAPHY..................................................................................................................... 79
5.1 PETROGRAPHY OF THE ANORTHOSITIC ROCK SUITE .............................................................. 79
5.1.1 Igneous textures ........................................................................................................... 79
5.1.1.1 White anorthosite suite.......................................................................................... 79
5.1.1.2 Dark anorthosite suite ........................................................................................... 82
5.1.1.2.1 Pyroxene-bearing anorthosite and leucogabbronorite.................................... 83
5.1.1.2.2 Olivine-bearing anorthosite and leucotroctolite............................................. 85
5.1.2 Structures of subsolidus origin .................................................................................... 88
5.1.2.1 Type I corona ........................................................................................................ 88
5.1.2.2 Type II corona....................................................................................................... 89
5.1.2.3 Ilmenite-magnetite-orthopyroxene symplectites................................................... 91
5.2 PETROGRAPHY OF THE FELSIC ROCK SUITE ........................................................................... 95
5.2.1 Granite and quartz-syenite........................................................................................... 96
5.2.2 Syenite .......................................................................................................................... 98
6 MINERAL CHEMISTRY..................................................................................................... 102
6.1 FELDSPAR .......................................................................................................................... 102
6.1.1 Anorthositic suite ....................................................................................................... 102
6.1.2 Felsic suite ................................................................................................................. 104
6.2 OLIVINE ............................................................................................................................. 106
6.3 PYROXENE ......................................................................................................................... 107
6.3.1 Orthopyroxene ........................................................................................................... 108
6.3.2 Clinopyroxene ............................................................................................................ 111
6.4 AMPHIBOLE........................................................................................................................ 111
6.5 BIOTITE .............................................................................................................................. 113
Table of contents
3
6.6 TITANITE ............................................................................................................................ 115
6.7 FE-TI OXIDES ..................................................................................................................... 115
6.7.1 Ilmenite....................................................................................................................... 116
6.7.2 Magnetite.................................................................................................................... 116
6.7.3 Hematite ..................................................................................................................... 117
6.8 GARNET ............................................................................................................................. 117
6.9 EPIDOTE ............................................................................................................................. 119
6.10 CHLORITE ......................................................................................................................... 119
6.11 COMPARATIVE APPROACH ................................................................................................ 120
7 GEOCHEMISTRY ................................................................................................................ 122
7.1 MAJOR AND TRACE ELEMENT GEOCHEMISTRY ................................................................... 122
7.1.1 Anorthositic suite........................................................................................................ 122
7.1.2 Felsic suite.................................................................................................................. 128
7.2 REE GEOCHEMISTRY.......................................................................................................... 131
8 OXYGEN ISOTOPE CONSTRAINTS................................................................................135
8.1 ANORTHOSITIC SUITE ......................................................................................................... 135
8.2 FELSIC SUITE ...................................................................................................................... 137
9 PRESSURE-TEMPERATURE CONDITIONS.................................................................. 140
9.1 THERMOBAROMETRY OF THE ANORTHOSITIC ROCKS .......................................................... 140
9.1.1 Magmatic textures ......................................................................................................140
9.1.2 Coronitic mineral assemblages.................................................................................. 142
9.1.3 Symplectitic mineral assemblages..............................................................................148
9.2 THERMOBAROMETRY OF THE SYENITIC SUITE .................................................................... 148
9.3 P-T EVOLUTION OF THE KUNENE INTRUSIVE COMPLEX AND THE TEMPORALLY ASSOCIATED
FELSIC SUITE............................................................................................................................. 152
10 DISCUSSION ....................................................................................................................... 157
10.1 CONSTRAINTS ON THE SOURCE AND EVOLUTION OF THE ANORTHOSITIC AND SYENITIC
MELTS ...................................................................................................................................... 157
10.2 PETROGENETIC MODEL ..................................................................................................... 160
11 COMPARISON WITH OTHER PARTS OF THE KIC.................................................. 164
PART III: FENITIZING PROCESSES INDUCED BY
FERROCARBONATITE MAGMATISM...................................................... 165
12 INTRODUCTION................................................................................................................ 165
12.1 CARBONATITES AND FENITIZATION .................................................................................. 165
12.2 SODALITE - A REVIEW ...................................................................................................... 166
13 PETROGRAPHY.................................................................................................................171
13.1 FENITIZED ANORTHOSITE ................................................................................................. 171
13.2 FENITIZED SYENITE .......................................................................................................... 177
13.3 FENITIZED NEPHELINE SYENITE ........................................................................................ 179
13.4 FENITIZED ULTRAMAFIC XENOLITHS ................................................................................ 182
13.5 CARBONATITIC BRECCIA .................................................................................................. 183
13.5.1 Massive carbonatitic breccia ................................................................................... 184
13.5.2 Layered carbonatitic breccia ................................................................................... 186
13.5.3 Sodalite-rich carbonatitic breccia ...........................................................................186
4
Table of contents
13.5.3.1 REE-poor samples............................................................................................. 186
13.5.3.2 REE-rich samples.............................................................................................. 192
13.5.4 Crystallization sequence of the carbonatitic breccia............................................... 194
13.6 FERROCARBONATITE VEINS.............................................................................................. 195
13.7 K-FELDSPAR-BIOTITE-ILMENITE-CALCITE ROCK .............................................................. 197
14 MINERAL CHEMISTRY................................................................................................... 198
14.1 CARBONATE MINERALS .................................................................................................... 198
14.1.1 Ankerite-dolomite solid solution .............................................................................. 198
14.1.2 Calcite ...................................................................................................................... 204
14.1.3 Siderite-magnesite solid solution ............................................................................. 206
14.1.3 Carbocernaite .......................................................................................................... 207
14.1.4 Strontianite............................................................................................................... 208
14.2 FELDSPAR ........................................................................................................................ 208
14.2.1 Potassium feldspar................................................................................................... 208
14.2.2 Plagioclase............................................................................................................... 209
14.3 NEPHELINE ....................................................................................................................... 211
14.4 SODALITE ......................................................................................................................... 212
14.5 CANCRINITE GROUP ......................................................................................................... 215
14.6 ANALCITE ........................................................................................................................ 216
14.7 AEGIRINE ......................................................................................................................... 217
14.8 AMPHIBOLE...................................................................................................................... 218
14.9 WHITE MICA..................................................................................................................... 219
14.10 BIOTITE .......................................................................................................................... 221
14.11 APATITE ......................................................................................................................... 227
14.12 PYROCHLORE ................................................................................................................. 229
14.13 FE-TI OXIDES ................................................................................................................. 230
14.13.1 Ilmenite................................................................................................................... 230
14.13.2 Magnetite ............................................................................................................... 231
14.13.3 Hematite ................................................................................................................. 232
14.13.4 Rutile ...................................................................................................................... 232
14.14 SULPHIDES ..................................................................................................................... 233
14.14.1 Pyrite...................................................................................................................... 233
14.14.2 Chalcopyrite........................................................................................................... 235
14.14.3 Pyrrhotite ............................................................................................................... 235
14.14.4 Pentlandite ............................................................................................................. 236
14.14.5 Bornite.................................................................................................................... 236
14.14.6 Chalcocite-digenite ................................................................................................ 236
14.14.7 Millerite.................................................................................................................. 237
14.14.8 Siegenite ................................................................................................................. 239
14.14.9 Violarite-polydymite............................................................................................... 239
14.14.10 Fletcherite ............................................................................................................ 241
14.15 BARITE ........................................................................................................................... 242
14.16 UNSPECIFIED NA-AL PHASE ........................................................................................... 242
14.17 TRACE AND RARE EARTH ELEMENT CHEMISTRY OF SELECTED MINERALS – A COMBINED
LA-ICPMS – SRXRF STUDY .................................................................................................. 245
14.17.1 Carbonates ............................................................................................................. 245
14.17.2 Sodalite................................................................................................................... 248
14.17.3 Biotite ..................................................................................................................... 249
14.17.4 Apatite .................................................................................................................... 249
14.17.5 Pyrochlore.............................................................................................................. 252
Table of contents
5
15 GEOCHEMISTRY .............................................................................................................. 255
15.1 MAJOR AND TRACE ELEMENT GEOCHEMISTRY ................................................................. 255
15.1.1 Fenitized anorthosite................................................................................................ 256
15.1.2 Fenitized syenite.......................................................................................................259
15.1.3 Fenitized nepheline syenite ......................................................................................261
15.1.4 Carbonatites ............................................................................................................. 262
15.1.4.1 Carbonatitic breccia...........................................................................................262
15.1.4.2 Ferrocarbonatite veins ....................................................................................... 267
15.2 RARE EARTH ELEMENT GEOCHEMISTRY ........................................................................... 268
15.3 DISCUSSION AND INTERPRETATION OF THE CARBONATITE EVOLUTION ............................ 271
15.4 COMPARISON WITH THE LUPONGOLA CARBONATITES ...................................................... 272
16 MASS BALANCE CALCULATIONS ............................................................................... 273
16.1 THE GRESENS APPROACH ................................................................................................. 273
16.2 MASS CHANGE CALCULATIONS FOR THE FENITIZATION OF ANORTHOSITES AND SYENITES274
16.3 ORIGIN AND NATURE OF THE FENITIZING SOLUTIONS ....................................................... 279
17 FLUID INCLUSION STUDIES.......................................................................................... 282
17.1 MICROTHERMOMETRIC STUDIES ....................................................................................... 282
17.1.1 H2O-rich inclusions.................................................................................................. 283
17.1.1.1 Fluid inclusions in sodalite................................................................................ 283
17.1.1.2 Fluid inclusions in carbonate............................................................................. 285
17.1.1.3 Fluid inclusions in quartz .................................................................................. 286
17.1.2 CO2-rich inclusions .................................................................................................. 287
17.1.3 Discussion and interpretation .................................................................................. 288
17.2 HIGH-RESOLUTION SYNCHROTRON MICRO-XRF ANALYSIS .............................................. 290
17.3 IMPLICATIONS FOR THE EVOLUTION OF THE FENITIZING FLUIDS ....................................... 293
18 STABLE ISOTOPE COMPOSITION ............................................................................... 295
18.1 OXYGEN AND CARBON ISOTOPIC COMPOSITION .............................................................. 295
18.1.1 Oxygen isotopic composition of anorthositic rocks .................................................295
18.1.2 Oxygen isotopic composition of fenitized syenites and nepheline syenites ..............296
18.1.3 Oxygen and carbon isotopic compositions of the carbonatitic breccia and the
ferrocarbonatite veins .........................................................................................................297
18.2 SULPHUR ISOTOPIC COMPOSITION..................................................................................... 300
19 P-T EVOLUTION OF THE CARBONATITIC MELTS AND FLUIDS ....................... 304
19.1 EMPLACEMENT AND CRYSTALLIZATION OF THE CARBONATITES ...................................... 304
19.2.1 Biotite-apatite thermometry and HF barometry ......................................................305
19.2.2 Calcite-dolomite thermometry..................................................................................309
19.1.3 Ankerite-magnetite oxygen isotope equilibria.......................................................... 309
19.1.4 Pyrite-chalcopyrite sulphur isotope equilibria ........................................................ 311
19.1.5 Stability relations of Co-Ni-Fe sulphides................................................................. 311
19.2 EMPLACEMENT AND FENITIZATION OF THE NEPHELINE SYENITE ...................................... 314
19.3 FENITIZATION OF ANORTHOSITIC XENOLITHS AND SYENITES– IMPLICATIONS FROM FLUIDINCLUSION ISOCHORE DATA ...................................................................................................... 316
20 CONCLUSIONS................................................................................................................... 319
20.1 ORIGIN AND EVOLUTION OF THE SWARTBOOISDRIF CARBONATITES ................................. 319
20.2 THE BREAKDOWN OF IGNEOUS MINERALS DURING FENITIZATION..................................... 322
20.3 LINKS TO OTHER CARBONATITE COMPLEXES OF NW NAMIBIA AND SW ANGOLA ........... 323
20.4 CONCLUDING REMARKS ................................................................................................... 324
6
Table of contents
21 REFERENCES..................................................................................................................... 325
Zusammenfassung
7
ZUSAMMENFASSUNG
Die mesoproterozoische Entwicklung Namibias ist durch wiederholte magmatische Aktivität
gekennzeichnet, welche zur Intrusion verschiedener Gesteinsschmelzen führte. Zunächst erfolgte die
Platznahme der anorthositischen Magmen des Kunene-Intrusiv-Komplexes (KIK), der mit einer Fläche
von etwa 18000 km2 das vermutlich größte Anorthosit-Massiv der Welt darstellt. Die Intrusion der
anorthositischen Magmen erfolgte vor etwa 1385-1347 Ma innerhalb der amphibolit- bis granulitfaziell
metamorphen Gesteine des paläo- bis meso-proterozoischen Epupa-Komplexes am Südwestrand des
Kongo-Kratons. Der KIK, ein typischer massif-type Anorthosit-Komplex, ist in Form einer N-S
ausgelängten Intrusion über 300 km Länge und 30-80 km Breite im Südwesten Angolas und
untergeordnet im Nordwesten Namibias aufgeschlossen. Er kann grob in zwei Haupteinheiten unterteilt
werden, die durch einen Gürtel granitischer Intrusionen (1400-1300 Ma) getrennt sind: (1) Ein Grossteil
des angolanischen KIK zeigt eine ausgeprägte N-S Orientierung und setzt sich überwiegend aus schwach
alterierten, dunkel gefärbten Anorthositen, Leukogabbros und Leukotroktolithen zusammen. (2) Dagegen
zeigt der südliche Teil des KIK, inklusive des namibischen Teiles, eine deutliche E-W Orientierung. In
dieser Region dominiert eine ältere, vollständig alterierte und weiß gefärbte Anorthosit-Varietät, der
sogenannte weiße Anorthosit. Dieser wurde entlang von etwa E-W-streichenden, extensionsgebundenen
Störungen von einer zweiten Anorthosit-Generation durchschlagen, die heute in Form von E-Wstreichenden Höhenrücken aus den Ebenen von weißem Anorthosit herauspräpariert ist. Dieser
kogenetisch entstandene Anorthosit-Typ entspricht in seinem Auftreten und seiner Mineralogie den
Anorthositen der nördlichen Einheit und wird aufgrund seiner dunkelgrauen bis schwarzen Färbung als
dunkler Anorthosit bezeichnet. Der Anorthosit-Komplex wurde nahe seiner südöstlichen Begrenzung, im
Bereich der Siedlung Swartbooisdrif, von zahlreichen extensionsgebundenen Störungen und Scherzonen
durchschlagen. In diese SE-NW und untergeordnet ENE-WSW streichenden Schwächezonen intrudierten
Quarz-Syenite und Syenite, die ein Alter von etwa 1380-1340 Ma aufweisen, sowie jüngere ankeritische
Karbonatite (ca. 1140-1120 Ma) und Nephelin-Syenite. Metasomatische Wechselwirkungen zwischen
Fluiden der karbonatitischen Schmelzen und intensiv zerscherten Syeniten sowie Anorthositen führten
innerhalb der Störungen zur Bildung ökonomisch bedeutsamer Sodalith-Vorkommen, die auf ein Areal
von etwa 100 km2, im Bereich der Siedlung Swartbooisdrif, begrenzt sind.
Ziel der mineralogisch-petrologisch-geochemischen Untersuchungen war eine vielfältige und
vergleichende Analytik möglichst vieler Proben unterschiedlicher Lithologie. Diese wurden sowohl an
den Anorthositen und Syeniten als auch an den jüngeren Karbonatiten und Nephelin-Syeniten
durchgeführt und umfassten die petrographische Bearbeitung von Dünnschliff-Präparaten und deren
detaillierte Mikrosonden-analytische Bearbeitung, geochemische Untersuchung der GesamtgesteinHaupt- und Spurenelement-Analytik und der Gesamtgestein-Seltenerdelement-Analytik ausgesuchter
Proben, mikrothermometrische Untersuchungen sowie Synchrotron-XRF-Untersuchungen von FluidEinschlüssen, Laser-ICPMS und Synchrotron-XRF Analysen der Spurenelement-Gehalte ausgesuchter
Minerale sowie O-, C- und S-Isotopenanalytik. Hierdurch sollte eine umfassende Vorstellung über die
genetische und zeitliche Stellung der verschiedenen Gesteinstypen und insbesondere über die
mineralbildenden Vorgänge gewonnen werden, die zur Abscheidung der metasomatisch gebildeten
Mineralparagenesen geführt haben.
8
Zusammenfassung
Die Anorthosite des KIK und die assoziierten felsischen Gesteinseinheiten: Für den weißen und
den dunklen Anorthosit konnten anhand der Feldbefunde, des Mineralbestandes und Mineralchemismus,
des Gefüges sowie des Gesamtgesteinschemismus zwei zeitlich getrennte aber kogenetische magmatische
Ereignisse charakterisiert werden. Die typische primär-magmatische Mineralogie in den Gesteinen des
KIK umfasst: Plagioklas (weißer Anorthosit: An37-53; dunkler Anorthosit: An43-75) ± Olivin ±
Orthopyroxen ± Klinopyroxen + Fe-Ti-Oxide + Apatit ± Zirkon. Spätmagmatische Säume von Amphibol
(meist Magnesio-Hastingsit/Pargasit) und Biotit umgeben die Pyroxene und Ilmenit. Weiße Anorthosite
sind im Gegensatz zu den dunklen Anorthositen durch eine durchgreifende Sericitisierung,
Saussuritisierung und Albitisierung der Plagioklase und Alteration der Fe-Mg Silikate charakterisiert.
Geothermometrische Untersuchungen ergaben Temperaturen von 985-950°C für die spätmagmatische
Kristallisation von Amphibol. Merkmale einer Reequilibrierung der Anorthosite unter granulit- bis
amphibolitfaziellen Bedingungen sind in Leukotroktolithen zugegen. In diesen Gesteinen werden
magmatische Olivine durch Orthopyroxen-Granat-Amphibol-Koronen und Ilmenite durch Biotit-GranatKoronen umgeben. Anhand geothermobarometrischer Untersuchungen wurden Bedingungen von 760 ±
100°C und 7.3 ± 1 kbar für die Bildung der Koronen ermittelt, welche sich unter annähernd isobarer
Abkühlung bis zu Temperaturen von 680-630°C und 6.2 ± 0.4 kbar fortsetzte. In Kombination mit den
Kristallisationsbedingungen der spätmagmatischen Amphibole resultiert für die Anorthosite ein P-T-Pfad
der annähernd isobaren Abkühlung unter amphibolit- bis granulit-faziellen Bedingungen.
In den zeitgleich intrudierten Quarz-Syeniten und Syeniten kristallisierten zunächst K-Feldspat,
Plagioklas und Klinopyroxen und/oder Hastingsit, gefolgt von einer zweiten Generation von K-Feldspat,
spät-magmatischem Hastingsit und Fe-Ti-Oxiden. Letztere werden vereinzelt durch schmale TitanitSäume umgeben. In Folge einer Subsolidus-Reequilibrierung der Syenite wurde Hastingsit partiell durch
Hastingsit-Ilmenit-Magnetit-Plagioklas-Symplektite ersetzt. Die magmatische Kristallisation von KFeldspat fand unter Temperaturbedingungen von 890-790°C statt, welche vermutlich nahe des Liquidus
der syenitischen Schmelzen liegen. Für die spät-magmatische Kristallisation von Hastingsit wurden
Drucke von 6.5 ± 0.6 kbar ermittelt, welche gut mit den für die Anorthosite kalkulierten Drucken
übereinstimmen. Die späte Bildung von Ilmenit-Magnetit-Plagioklas-Symplektiten erfolgte unter
ähnlichen Druckbedingungen (6.6-6.9 kbar), welche auf eine annähernd isobare Abkühlung der Syenite
hinweisen.
Mit dem Ziel, die Natur der Magmenquelle der Anorthosite und Syenite zu charakterisieren und
die Frage zu klären, ob Syenite und Anorthosite eventuell aus einem gemeinsamen Stamm-Magma
hervorgegangen sind, wurden geochemische Untersuchungen durchgeführt, sowie die Sauerstoff-Isotopie
von Feldspat-Separaten bestimmt. Die Ergebnisse dieser Untersuchungen belegen, dass es die StammMagmen der Anorthosite des KIK stark fraktionierte, basaltische Magmen darstellen. Die gewonnenen
Sauerstoff-Isotopendaten für Plagioklas (5.61-6.13 ‰ δ18O) legen zudem nahe, dass es sich hierbei um
Schmelzen handelt, die durch partielle Aufschmelzung des Mantels entstanden sind. Ähnliche Werte
wurden für Plagioklas-Separate aus Proben des durchgreifend alterierten weißen Anorthosits ermittelt
(5.35-6.14‰). Die gute Übereinstimmung der δ18OSMOW-Werte von Plagioklasen weißer und dunkler
Anorthosite lässt darauf schließen, dass es sich bei den Fluiden, die zur Überprägung des weißen
Anorthosits führten, um magmatische Fluide anorthositischer Schmelzen handelte. Petrographische und
geochemische Untersuchungen bestätigen, dass nahezu alle Anorthosite im Zuge ihres Aufstiegs eine
variable Kontamination durch krustales Material erfuhren, die sich am stärksten auf die früh intrudierten
Anorthosit-Magmen auswirkte. Die für die Syenite ermittelten Ergebnisse belegen eindeutig, dass es sich
Zusammenfassung
9
bei Anorthositen und Syeniten um chemisch unabhängige Systeme handelt, da (1) chemische Kontinuität
zwischen Anorthositen und Syeniten nicht beobachtet wurde, (2) die Spurenelementgehalte der felsischen
Gesteine auf eine krustale oder zumindest eine gemischte Kruste-Mantel-Quelle hindeuten, (3) Chondritnormierte Seltenerd-Element-Muster der Syenite keine negative Eu-Anomalie aufweisen, wie sie von
Fraktionierungsprodukten von Schmelzen, welche zuvor in großen Mengen Plagioklas kristallisierten,
erwartet werden könnte und (4) die δ18OSMOW-Werte von magmatisch kistallisiertem Feldspat der Syenite
mit 7.20-7.92 ‰ etwa 1.6 ‰ höher liegen als die der magmatisch gebildeten Plagioklase der Anorthosite.
Dementsprechend sind die Anorthosite und Syenite zwar zeitgleich intrudiert, eine Entwicklung aus nur
einem Stamm-Magma kann jedoch eindeutig ausgeschlossen werden. Vielmehr liegt ihre enge räumliche
und zeitliche Assoziation darin begründet, dass der Aufstieg und die Platznahme der Stamm-Magmen der
Anorthosite zum partiellen Aufschmelzen der Unterkruste und somit zur Bildung potentieller StammMagmen der Syenite führten.
Die Karbonatite von Swartbooisdrif: Im Verlauf eines weiteren, jüngeren magmatischen
Ereignisses erfolgte vor etwa 1140-1120 Ma zunächst die Platznahme einer ersten Karbonatit-Generation,
welche in erheblichem Maße durch fragmentiertes, anorthositisches und syenitisches Gesteinsmaterial
kontaminiert ist und daher entsprechend ihres makroskopischen Erscheinungsbildes als karbonatitische
Brekzie bezeichnet wird. Diese wird von zahlreichen jüngeren und geringmächtigen FerrokarbonatitAdern durchschlagen, die sich durch eine nur geringe Kontamination mit Nebengesteinsmaterial
auszeichnen.
Die
gewonnenen
Ergebnisse
der
mineralogisch-petrologisch-geochemischen
Untersuchungen der Karbonatite belegen eindeutig, dass die Karbonatiten beider Generationen
magmatische Kristallisationsprodukte fraktionierter Mantelschmelzen und nicht hydrothermal gebildete
Gesteinseinheiten darstellen. Die Karbonatite treten als Gänge und Adern auf, welche die ältere
Gesteinseinheiten durchschlagen. Syenite und Anorthosite haben im Kontakt zu den Karbonatiten eine
metasomatische Überprägung erfahren; Nebengesteinsklasten sind in der Fliessrichtung der Karbonatite
orientiert, zum Teil rotiert und werden von Karbonat-reichen Lagen umflossen. Diese Annahme wird
durch die Mineralogie beider Karbonatit-Generationen, i.e. (1) Ankerit + Kalzit + Magnetit + Biotit ±
Ilmenit ± Apatit ± Pyrochlor ± Sulfide in der Karbonat-Matrix der karbonatitischen Brekzie und (2)
Ankerit + Kalzit + Magnetit ± Pyrochlor ± Rutil der jüngeren, nahezu Xenolith-freien FerrokarbonatitAdern, ihren geochemischen Signaturen sowie durch die Sauerstoff- und Kohlenstoff-Isotopie separierter
Ankerite (8.91-9.73 ‰ δ18O und –6.73 bis –6.98 ‰ δ13C) bestätigt, wobei allerdings die O-Isotopendaten
eine schwache hydrothermale Alteration beider Karbonatit-Typen nahe legen. Die Karbonatite beider
Generationen können somit als Kristallisationsprodukte stark fraktionierter Mantelschmelzen mit
ferrokarbonatitischer Zusammensetzung charakterisiert werden. Das Auftreten von Biotit, Apatit, Ilmenit
und Sulfiden in der karbonatitischen Brekzie deutet darauf hin, dass diese Schmelzen durch
vergleichsweise hohe Gehalte an Si, Al, K, Ti, S und P gekennzeichnet waren, was über geochemische
Untersuchungen bestätigt werden konnte. Dagegen wurden diese Minerale in den jüngeren
Ferrokarbonatit-Adern nicht beobachtet, welche zudem vergleichsweise geringe Gehalte an Fe, Mg, Si,
Al, Ti, K, S und P aufweisen. Diese Ergebnisse legen nahe, dass die Bildung der späten Ferrokarbonatite
im Anschluss an die Kristallisation von Biotit, Apatit, Ankerit, Ilmenit, Magnetit und den Sulfiden
erfolgte. Die jüngeren Ferrokarbonatite stellen somit die Fraktionierungsprodukte der für die Bildung der
karbonatitischen Brekzie verantwortlichen Karbonatit-Magmen dar. Allerdings belegen die
geochemischen Signaturen der karbonatitischen Brekzie, dass diese Fraktionierungs-Prozesse bereits vor
10
Zusammenfassung
der Intrusion der älteren Karbonatit-Einheit aktiv waren. Im Gegensatz hierzu können extreme Sr- und
SEE-Anreicherungen zweier Proben der karbonatitischen Brekzie nicht über extreme Fraktionierung der
Karbonatit-Schmelzen abgeleitet werden. In diesen Proben werden magmatische Ankerit-Kristalle
randlich durch das SEE-Karbonat Carbocernait ersetzt. Zudem füllen Baryt und Strontianit Zwickel
zwischen einzelnen Ankerit-Körnern. Diese Mineral-Umwandlungen und Neubildungen resultieren aus
einer hydrothermalen Überprägung der Gesteine durch späte, Sr- und LSEE-reiche karbonatitische Fluide.
Prozesse der Fenitisierung: Feldbefunde in der Region Swartbooisdrif deuten darauf hin, dass die
Devolatisierung früh intrudierender karbonatitischer Magmen zur metasomatischen Überprägung intensiv
gescherter Anorthosite und Syenite im Kontakt zu den Karbonatit-Gängen führte. Zudem wirkte sich
diese Fenitisierung auch auf die in den Karbonatiten eingeschlossenen Nebengesteinsbruchstücke aus. Im
Zuge dieses Ereignisses wurden insbesondere Plagioklas und K-Feldspat der Anorthosite und Syenite
sowie Nephelin eines Nephelin-Syenites in Albit und/oder Sodalith umgewandelt. Unter Verwendung
konventioneller Geothermometrie in Kombination mit den für Fluid-Einschlüsse kalkulierten Isochoren
konnten die Druck- und Temperaturbedingungen sowohl der Platznahme der Karbonatite als auch der im
Zuge der Fenitisierung initiierten mineralum- und -neubildenden Prozesse bestimmt werden. Hierbei
wurden die Entwicklung sowie die Zusammensetzung der fenitisierenden Fluide aus der zeitlichen
Abfolge der Mineral-Reaktionen abgeschätzt. Quantitative Daten wurden über mikrothermometrische
Untersuchungen und Synchrotron-XRF-Analysen von Fluid-Einschlüssen gewonnen, die in
verschiedenen, metasomatisch neugebildeten und überprägten Mineralen eingeschlossen sind. Die hierbei
gewonnenen Ergebnisse lassen folgende Rückschlüsse zu: (1) Im Zuge eines frühen metasomatischen
Ereignisses erfolgte die Umbildung von Plagioklas der Anorthosite und K-Feldspat der Syenite in nahezu
reinen Albit. Zeitgleich wurden wasserfreie Fe-Mg-Silikate der Anorthosite und Syenite in Biotit
umgewandelt. Wie texturelle Beobachtungen belegen, fanden diese Prozesse bereits vor der Intrusion der
Karbonatit-Magmen statt. Die metasomatischen Reaktionen implizieren, dass es sich bei den
verantwortlichen Fluiden um Na-reiche, wässrige Lösungen handelte. (2) Im Anschluss an die
Platznahme der karbonatitischen Brekzie in ca. 12-15 km Tiefe (4-5 kbar), erfolgte die Kristallisation von
Apatit und Biotit über ein Temperaturintervall von 1200-630°C. Etwa zeitgleich kristallisierten die
wasserfreien Minerale Ankerit und Magnetit, was eine erneute Wassersättigung des Karbonatit-Magmas
zur Folge hatte. Unter Temperaturen von 800-530°C bewirkte die Zirkulation NaCl-reicher wässriger
Fluide (19-30 Gew.% NaCl äquivalent) die Umwandlung von Albit der eingeschlossenen
Nebengesteinsbruchstücke in Sodalith. Unter ähnlichen Temperaturen (ca. 775-700°C) wurde Nephelin
eines angrenzenden Nephelin-Syenites zu Sodalith-Muskovit-Paragenesen abgebaut. Wie SRXRFAnalysen belegen, enthielten die fenitisierenden Fluide zudem geringe Konzentrationen an Sr, Ba, Fe, Nb
und SEE. Ein großer Teil des in den Fluiden enthaltenen Natriums wurde während dieser Ereignisse
aufgebraucht. Dementsprechend weisen sekundäre Fluid-Einschlüsse eines benachbarten Quarz-Syenit
Ganges geringere Salinitäten (9-24 Gew.% NaCl äquivalent) und zugleich höhere Gehalte an K, Ca, Fe,
Ti, Sr und Ba auf. Die hohen modalen Anteile an dem auf Kosten von Plagioklas neugebildeten Sodalith
in der karbonatitischen Brekzie legen nahe, dass das Karbonatit-Magma vor der Fenitisierung erhebliche
Gehalte an Natrium enthielt. Im Gegensatz zur karbonatitischen Brekzie konnten jedoch im Kontakt zu
den jüngeren Ferrokarbonatit-Adern keine Fenit-Aureolen beobachtet werden; die Schmelzen scheinen
dementsprechend ein nur geringes metasomatisches Potential besessen zu haben. (3) Während eines
späten Stadiums der Fenitisierung bildeten sich Cancrinit und Muskovit auf Kosten von Albit der im
Karbonatit eingeschlossenen Nebengesteinsbruchstücke. Es kann daher davon ausgegangen werden, dass
es sich bei den fenitisierenden Fluiden noch immer um wässrige Lösungen handelte, die jedoch nicht
Zusammenfassung
11
mehr ausreichend Natrium enthielten, um Sodalith zu bilden. Die Si-freie Zusammensetzung einer bislang
unbekannten Na-Al-Phase, welche sich später auf Kosten von Sodalith bildete, legt zudem nahe, dass
auch späte fenitisierende Fluide stark Si-untersättigt waren.
Die vorliegende Arbeit zeigt, dass die Bildung der Ferrokarbonatite sowie die von ihnen
verursachte Kontakt-Metasomatose das Resultat zahlreicher, miteinander eng verknüpfter magmatischer
und metasomatischer Prozesse ist. Nahezu jeder dieser Prozesse kann mit spezifischen Mineralum- und neubildenden Vorgängen in Zusammenhang gebracht werden, die letztendlich zur Überprägung der
Anorthosite und Syenite zu Albit- und Sodalith-reichen Feniten führte.
12
Abstract
ABSTRACT
During the Mesoproterozoic large volumes of magma were repeatedly emplaced within the
Palaeoproterozoic basement of north-west Namibia. Magmatic activity started with the intrusion of the
anorthositic rocks of the Kunene Intrusive Complex at 1,385-1,347 Ma which, with a total surface area of
18,000 km2, is one of the largest anorthosite bodies of the world. At its south-eastern margin the Kunene
Intrusive Complex was subsequently invaded by Mesoproterozoic quartz-syenite and syenite dykes
(1,380-1,340 Ma) as well as by younger carbonatites (1,140-1,120 Ma) and nepheline syenites along ENE
and SE trending faults. Older carbonatite intrusions, designated as carbonatitic breccia, frequently contain
anorthositic and syenitic wallrock fragments set in a ferrocarbonatite matrix, whereas subordinate veins of
massive ferrocarbonatite are almost xenolith-free and cut through the main carbonatite dykes.
Metasomatic interaction between carbonatite-derived fluids and both neighbouring lithologies and
incorporated anorthosite xenoliths led to the formation of economically important sodalite occurrences,
confined to an area of about 100 km2.
Investigated anorthosite samples display the typical magmatic mineral assemblage of plagioclase
(An37-75) ± olivine ± orthopyroxene ± clinopyroxene + ilmenite + magnetite + apatite ± zircon. Ilmenite
and pyroxene are surrounded by narrow reaction rims of biotite as well as pargasite. During the
subsolidus stage sporadic coronitic garnet-orthopyroxene-quartz assemblages around olivine and garnetbiotite assemblages around ilmenite were produced. Thermobarometry studies on amphiboles yielded
temperatures of 985-950°C for the late-magmatic crystallization of the anorthositic rocks. The chemical
compositions of garnet-orthopyroxene coronas and their textural relationships indicate, that the
emplacement and crystallization of the anorthosites was followed by subsolidus re-equilibration of the
plutonic body at conditions of 760 ± 100°C and 7.3 ± 1 kbar. Near-isobaric cooling proceeded to
temperatures 680-630°C and pressures of 6.2 ± 0.4 kbar.
In the contemporaneously emplaced quartz-syenites and syenites potassium feldspar, plagioclase,
hastingsite and/or clinopyroxene crystallized first, followed by a second generation of potassium feldspar
as well as interstitial hastingsite and Fe-Ti oxides, which may in turn be surrounded by narrow rims of
titanite. During the subsolidus re-equilibration of the syenites interstitial hastingsite was partially replaced
by hastingsite-ilmenite-magnetite-plagioclase symplectites. Magmatic crystallization of potassium
feldspar in the syenites occurred under temperatures of 890-790°C, which are presumably near the
liquidus of the rocks. For the late-magmatic crystallization of hastingsite pressures of 6.5 ± 0.6 kbar have
been obtained, which are similar to those derived for late-stage garnet-orthopyroxene coronas around
olivine in the anorthosites. For the subsolidus formation of hastingsite-ilmenite-magnetite-plagioclase
symplectites pressures of 6.6-6.9 kbar have been determined, implying that the syenite emplacement was
followed by almost isobaric cooling.
In order to constrain the source rocks of the two suites, oxygen isotope analyses of feldspar from
the anorthosites and syenites as well as geochemical bulk rock analyses were carried out. In case of the
Kunene Intrusive Complex, the general geochemical characteristics of all anorthositic lithologies are in
excellent agreement with a derivation from fractionated basaltic liquids, with the δ18O values obtained for
magmatic plagioclase (5.88 ± 0.19 ‰ δ18O) proving their derivation from mantle-derived magmas.
However, the mineralogical and isotopic data suggest that all studied anorthosite samples underwent a
minor degree of crustal contamination, which was most prominent during the early intrusion stages. The
Abstract
13
results obtained for the felsic suite, in turn, provide convincing evidence against consanguinity of the
anorthosites and the syenites, i.e. (1) compositional gaps are developed between the major and trace
element data of the two suites; chemical continuity is lacking, (2) trace element data of the felsic suite
points to a crustal or at least a mixed crustal-mantle source of the felsic suite, (3) syenites do not exhibit
ubiquitous negative Eu-anomalies in their REE patterns, which would be expected from fractionation
products of melts that previously formed extensive plagioclase cumulates and (4) magmatic feldspar δ18O
values from both the quartz-syenite and the syenites fall in a restricted range of 7.20-7.92 ‰, which,
however, are about 1.6 ‰ higher than the average magmatic plagioclase δ18O of the anorthosites.
Conformably, the crustal-derived felsic and the mantle-derived anorthositic suite are suggested to be
coeval but not consanguineous. Their spatial and temporal association can be accounted for, if the heat
necessary for crustal melting is provided by the upwelling and emplacement of mantle-derived melts,
parental to the anorthosites.
In order to constrain the source of the 1,140-1,120 Ma carbonatites and to elucidate the fenitizing
processes, which led to the formation of the sodalite occurrences of Swartbooisdrif, detailed
mineralogical and geochemical investigations, stable isotope (C,O,S) analyses and fluid inclusion
measurements (microthermometrical studies and synchrotron-micro-XRF analyses) of both the
carbonatites and the fenitized lithologies have been combined.
There is striking evidence that carbonatites of both generations are magmatic rather than
hydrothermal in origin. They occur as dykes and veins with cross-cutting relationships and margins
disturbed by fenitic aureoles, and contain abundant flow-oriented xenoliths from the surrounding
anorthosite and syenite. The mineral assemblage of both carbonatite generations of ankerite + calcite +
ilmenite + magnetite + biotite ± apatite ± pyrochlore ± sulphides in the main carbonatite body and
ankerite + calcite + magnetite ± pyrochlore ± rutile in the younger and almost silicate-free
ferrocarbonatite veins, their geochemical characteristics and the O and C isotope values (8.91 to 9.73 and
–6.73 to –6.98, respectively) again indicate igneous derivation, with the δ18O values suggesting only
minor subsolidus alteration. Carbonatites of both generations can hence be described as mantle-derived,
magmatic ferrocarbonatite magmas. The presence of apatite and biotite phenocrysts in the carbonatitic
breccia is suggestive of a ferrocarbonatite melt that additionally contained minor amounts of P, K, Si and
Al. The absence of these phases in the ferrocarbonatite veins implies that the ferrocarbonatite veins
represent late-stage fractionation products of the ferrocarbonatite magma which formed the carbonatitic
breccia, with their emplacement post-dating the extensive crystallization of biotite, ankerite, Fe-Ti oxides,
apatite, and sulphides. However, the geochemical characteristics of the carbonatitic breccia itself provides
evidence that these fractionation processes were already active previous to the emplacement of the older
carbonatite unit. In contrast the extreme REE-enrichment, observed in two samples of the carbonatitic
breccia, does not result from fractionation of the carbonatite melt, but is caused by a hydrothermal
overprint by late-stage and highly evolved carbonatite fluids, which contained appreciable amounts of Sr
and LREE.
NaCl-rich fluids, released from the early intruding carbonatite melt caused the fenitization of
both, the incorporated wallrock xenoliths and the bordering anorthosite, syenite and nepheline syenite.
This process is mainly characterized by the progressive transformation of Ca-rich plagioclase, K-feldspar
and nepheline into albite and/or sodalite. Applying conventional geothermobarometry combined with
fluid-inclusion isochore data, it was possible to reconstruct the P-T conditions for the carbonatite
14
Abstract
emplacement and crystallization and for several mineral-forming processes during metasomatism. The
composition and evolutionary trends of the fenitizing solution were estimated from both the sequence of
metasomatic reactions within wallrock xenoliths in the carbonatitic breccia and fluid inclusion data
(microthermometric measurements combined with synchrotron-micro-XRF analysis): (1) The formation
of almost pure albite at the expense of former plagioclase and alkali-feldspar from anorthosite and
syenite, respectively, and the transformation of anhydrous Fe-Mg-silicates and amphibole into biotite in
the respective rock types predated the emplacement of the carbonatite melts. These metasomatic reactions
require that Na and H2O were the dominant components of this fluid. (2) After its final emplacement in
the crust at c. 12-15 km (4-5 kbar), the carbonatite magma again reached water saturation following the
contemporaneous formation of apatite and biotite at temperatures of c. 1200-630°C and fractionation of
the unhydrous phases ankerite and magnetite. In a temperature range of 800-530°C the circulation of
NaCl-rich aqueous brines (19-30 wt.% NaCl eq.) caused the transformation of albite of the incorporated
anorthositic xenoliths into sodalite. The sodalitization of nepheline in a bordering nepheline syenite dyke
occurred under similar temperatures (c. 775-700°C). Moreover, these fluids were capable of transporting
minor amounts of Sr, Ba, Fe, Nb, and LREE. As most of the sodium contained in the fenitizing agents
was consumed during this major sodalite-forming event, fluids trapped in quartz of a neighbouring
quartz-syenite dyke display lower salinities of (9-24 wt.% NaCl eq.) but, at the same time, contain higher
amounts of K, Ca, Fe, Ti, Sr, and Ba. The abundance of metasomatically formed sodalite in most samples
of the carbonatitic breccia, formed at the expense of plagioclase from anorthosite xenoliths, suggests that
these melts contained appreciable amounts of sodium before alkali-loss during fenitization. In contrast,
late-stage ferrocarbonatite veins incorporate previously fenitized xenoliths but rarely show any wallrock
alteration related to their emplacement. (3) During late stages of fenitization, cancrinite and muscovite
replace albite from feldspar-rich xenoliths in the carbonatitic breccia, hence implying that the evolved
fluids were still aqueous brines but no longer contained sufficient Na to form sodalite. The Si-free
composition of a subsequently formed and yet unspecified Na-Al phase, grown at the expense of sodalite,
provides evidence that even the late-stage carbonatite fluids were strongly deficient in Si.
This study demonstrates that the generation of ferrocarbonatite magmas and the associated
contact metasomatism in the Swartbooisdrif area are the result of several magmatic and interrelated
metasomatic processes, each inducing characteristic mineral reactions in the anorthositic and syenitic
wall-rocks.
Acknowledgements
15
ACKNOWLEDGEMENTS
Special thanks are due to Professor Dr. M. Okrusch who initiated this project. His unrestricted
support and encouragement as well as his never ending readiness for debates and discussions
greatly contributed to the success of this project. To Professor Dr. V. Lorenz, I wish to express
my gratitude for his guidance and interest during the field investigations and for his critical
review of the manuscript.
Many thanks are due to Dr. habil. N. Cook (Geological Survey of Norway, Trondheim) and
Dipl.-Min. S. Littmann (GeoForschungsZentrum Potsdam) who provided the sample material of
1997 and gave me a first insight into the geology of NW Namibia.
I am especially grateful to Dr. V. von Seckendorff for his always friendly willingness for
discussions concerning the complex geology of the Swartbooisdrif area and his critical
comments on the various manuscripts. Dipl.-Geol. S. Brandt is thanked for challenging
discussions regarding the metamorphic Epupa Complex surrounding the investigated field area
and for providing selected anorthosite and granite samples. I am grateful to Professor Dr. R.
Klemd, Mrs. R. Baur and Dr. H. Brätz for their help and advise with the XRF- and LA-ICPMSmeasurements and to Dr. habil. U. Schüssler for his ideas to improve the analytical approach of
the EMP work. Thanks are also due to Dr. K. Hradil for her support with the XRDmeasurements and to Mr. P. Späthe for his thorough preparation of the polished sections. Mrs. A.
Kirchner is thanked for her always friendly administrative help.
I am greatly indebted to Professor Dr. J. Hoefs (Göttingen) for providing the oxygen- and
carbon-isotope data and helping with their interpretation. Thanks are extended to Professor Dr. J.
Erzinger, Dr. P. Dulski, Dr. E. Kramer, Dipl.-Min. S. Littmann and Mrs. S. Tonn
(GeoForschungsZentrum Potsdam) for providing the opportunity to carry out REE analyses at
the GeoForschungsZentrum Potsdam and their help and advice with the sample preparation and
REE analysis. Many thanks go to Dr. K. Rickers (HASYLAB, Hamburg) for her introduction to
and help with the Synchrotron-Micro-XRF analysis. I am indebted to Dr. T. Wagner (University
of Tübingen) for performing the sulfur-isotope measurements, compiling the diagram of
literature data for thiospinel compositions and for various valuable discussions. Dr. U. Hünken
(KFA Jülich, Außenstelle Potsdam) is thanked for stimulating discussions, the assistance during
the microthermometric measurements and the careful reading of the fluid inclusion chapter and
16
Acknowledgements
Dipl.-Min. A. Klimera (Fraunhofer ISC, Würzburg) for his help with the density determination at
the Fraunhofer ISC, Würzburg.
The research project benefited from the support of the Geological Survey of Namibia. I am
obliged to Dr. G. Schneider, Mr. V. Petzel, Mr. B. Rösener and Mr. R. Wackerle. The permission
of Namibia Blue Sodalite Ltd. Company for access to, and sampling in the Swartbooisdrif
sodalite mine and the kind assistance of Mr. F. Augustin and his son during the field studies is
highly appreciated.
Many thanks are addressed to my friends and to the colleagues of the Institute of Mineralogy,
Würzburg. Special thanks go to my family and to Sönke Brandt for their unrestricted
encouragement and support during my work.
The financial support of Deutsche Forschungsgemeinschaft (grant Ok 2/64-1) is gratefully
acknowledged.
1 Introduction
17
1 Introduction
1.1 Location of the study area
The study area is located at the north-western border of Namibia, approximately 780 km north of
the Namibian capital Windhoek (Fig.1.1). It is dominated by the anorthositic rocks of the
Kunene Intrusive Complex (KIC) that, with a length of 350 km and a width of 30-80 km, extends
from SW Angola into NW Namibia (Fig. 1.2). Towards the north and north-west the study area
is restricted by the river Kunene, which marks the natural border between Namibia and Angola.
The south-western and southern boundaries of the study area are formed by the metamorphic
basement rocks of the Epupa Complex (EC). Rock samples were taken from a total area of
approximately 2,800 km2.
1.2 Geological framework
1.2.1 The metamorphic Epupa Complex
Metamorphic basement rocks in NW Namibia include the metamorphosed volcano-sedimentary
sequence of the Epupa Complex (EC) which forms the southern extension of the Proterozoic
Congo Craton (also referred to as the Angola-Kasai craton; see e.g. Clifford, 1970). The
Namibian EC was subdivided into two units by Brandt et al. (2000, 2003; Fig. 1.3):
(1) Upper amphibolite-facies rocks of the Orue Unit are the dominant rock unit, mainly
consisting
of
migmatitic
granitic
orthogneisses
intruding
subordinate
migmatitic
metasedimentary and metavolcanic rocks. A direct contact between the Orue Unit and the KIC
has not been observed, since it has been intruded by dolerite and syenite dykes.
(2) To the south, about 10 km south of the KIC, a limited but well defined area of
ultrahigh-temperature granulite-facies rocks has been recognized by Brandt et al. (2000, 2001,
2003), termed Epembe Unit. In its central part, this unit consists of a volcano-sedimentary
sequence of interlayered felsic and mafic granulites, intercalated with subordinate migmatitic
metasedimentary granulites. Rare sapphirine-bearing lenses of Mg-Al-rich granulites were
observed in layers of migmatitic metagreywackes (Brandt et al., 2000, 2001, 2003). In contrast,
the north-western and eastern parts of the Epembe Unit are dominated by massive mafic
granulites. The Epembe Unit is separated from the Orue Unit by a sub-vertical E-W trending
fault, termed Otjitambi-Ehomba Line, which was partly intruded by younger and
unmetamorphosed gabbroic intrusions.
18
1 Introduction
Fig. 1.1: Geographic overview map of Namibia (location of the study area marked).
According to Carvalho & Alves (1990, 1993) the Epupa Complex extends from Namibia
to southern Angola, where it is designated as Schist, Quartzite and Amphibolite Complex (Fig.
1.2). In contrast, the country rock in the north-western Angolan part of the KIC is mainly
composed of weakly deformed equigranular biotite-granites, termed Regional Granite (e.g.
Carvalho & Alves, 1990) of yet unknown metamorphic grade. Towards the East, the KIC is
buried by Kalahari sediments.
1.2.2 The anorthositic Kunene Intrusive Complex
The EC is intruded by the anorthositic rocks of the KIC, that, with an estimated total surface area
of 18,000 km2, is one of the largest but at the same time one of the least known massif-type
1 Introduction
19
anorthosite bodies in the world (Ashwal & Twist, 1994). With a length of 350 km and a width of
30-80 km, the N-S elongated complex extends from SW Angola into NW Namibia (Fig. 1.2).
Fig. 1.2: Geologic overview map of the Kunene Intrusive Complex in SW Angola and NW Namibia. Map
simplified and modified after Carvalho & Alves (1990). Outlined area marks the study area.
Only 2,500 km2 of the KIC crop out on Namibian territory south of the Kunene River, building
the mass of the Zebra Mountains. As already proposed by Morais et al. (1998) and Drüppel et al.
(2001) the KIC may be subdivided into two major units, separated by a NE-SW trending belt of
several granitoid intrusions:
20
1 Introduction
(1) The majority of the Angolan part of the KIC displays a N-S elongation and is mainly
composed of fresh to weakly altered, dark coloured anorthosites, leucogabbros and
leucotroctolites with subordinate gabbros, norites and troctolites (e.g. Carvalho & Alves, 1990;
Silva, 1992; Ashwal & Twist, 1994; Morais et al., 1998). Less than 10 % of these anorthositic
rocks show igneous lamination with a pronounced NE-SW fabric, defined by a preferred
orientation of plagioclase laths (Simpson, 1970; Morais et al., 1998).
(2) In contrast, the southernmost part of the KIC, including the Zebra Mountains (Fig. 1.3),
has a pronounced E-W elongation. It mainly consists of whitish to greenish, heavily altered
anorthosites and leucogabbros (cf. Simpson, 1970; Köstlin, 1974; Menge, 1996, 1998), the white
anorthosite (about 60 %). In contrast, dark coloured, weakly altered anorthosites,
leucotroctolites, leucogabbros and leuconorites, the dark anorthosite, are subordinate. Both the
white and the dark anorthosite are characterized by a massive appearance, due to a homogeneous
structure without preferred orientation of the plagioclase crystals. Cumulate textures are
relatively widespread on a local scale; yet no evidence for extensive layering or igneous
lamination was recognized, in contrast to the interpretation of aerial photographs by Menge
(1996, 1998).
A conspicuous feature of the KIC are elliptic patches of Fe-Ti-ore concentrations
(geological map), consisting predominantly of titanomagnetite and ilmenite with minor Fe-Mgsilicates (olivine and/or orthopyroxene) and sulphides. So far, neither geophysical data nor drill
holes exist which might provide evidence for their exact three-dimensional shapes. These
magnetite-ilmenite ores are restricted to the anorthositic rocks of the KIC and thus most probably
resemble exsolution features of immiscible oxide and silicate.
A series of mafic and ultramafic intrusions of yet unknown age, the settings of which
infer a genetic relationship to the anorthosite, crop out within the EC near the southern margin of
the KIC (Fig. 1.3). Several individual mafic and ultramafic (dunite, pyroxenite, gabbro,
leucogabbro, norite, hyperite, hornblende-hyperite, serpentinite) stocks and pipes that
predominately intrude E-W and SE trending weakness zones have been mapped by Menge
(1996). These rocks, the total volume of which is small compared to the KIC, have been
generally assumed to be of an age comparable to that of the anorthosite, yet certain field
relationships and geochemical patterns (Köstlin, 1967; Carvalho & Alves, 1990; Menge, 1996,
1998) suggest the intrusions to post-date the anorthosite emplacement.
1 Introduction
21
Fig. 1.3: Geologic overview map of the southern part of the Kunene Intrusive Complex. Map simplified and
modified after Menge (1998).
1.2.3 The syenite-carbonatite-lamprophyre suites of Swartbooisdrif and Epembe
Close to the southern margin of the KIC, in the Epembe and Swartbooisdrif regions, both the
KIC and the EC are transsected by numerous SE and NE to ENE trending extensional faults and
shear zones that were in parts intruded by various lamprophyre, carbonatite, nepheline syenite
and syenite dykes (Fig. 1.3):
(1)
In the Epembe area this alkaline suite mainly comprises small plugs and dykes of
nepheline syenite and minor syenite intrusions emplaced within the metamorphic rocks of the
Epupa Complex. These rocks clearly post-date the basic satellite intrusions, as evidenced by
ultramafic xenoliths within the nepheline syenite (Littmann et al., 2003). Also assigned to the
same alkaline suite are E-W-striking calciocarbonatite, ferrocarbonatite and lamprophyre dykes
(Ferguson et al., 1975).
22
1 Introduction
(2)
A second swarm of NE and SE striking dykes is exposed within the anorthositic
rocks of the Kunene Intrusive Complex in the Swartbooisdrif area. This rock suite is distinct in
appearance from that of the Epembe area. Felsic intrusives are mainly represented by syenite,
which generally forms NE and SE striking dykes, crosscutting the anorthosite complex. A major
swarm of ankeritic carbonatite dykes with a predominant SE, and a subordinate NE direction
occurs in this area, dissecting both the syenite dykes and the anorthositic rocks of the KIC. These
ankerite-rich carbonatites contain abundant xenoliths of anorthosite and syenite, surrounded by a
laminated carbonatite matrix. Based on their macroscopic appearance these rocks are designated
as carbonatitic breccia, even though they are clearly not a carbonatitic breccia sensu stricto. The
carbonatitic breccia may contain variable amounts of sodalite, which are mined in an area of
approximately 100 km2, about 7.5 km NW of Swartbooisdrif (Fig. 1.3; Plate 1: geological map).
A second generation of almost silicate-free ferrocarbonatite veins, transsecting the carbonatitic
breccia, is present in minor proportions.
1.3 Available geochronological data
The available geochronological data is summarised in Table 1. The age of the EC has been
commonly regarded as Palaeo- to Mesoproterozoic (Martin 1965; Hedberg 1979). One date has
been constrained by conventional U-Pb dating on zircons of a syntectonically intruded granite at
Ruacana yielding an upper intercept age of 1,795 +33/-29 Ma (Tegtmeyer & Kröner, 1985).
Recently however, Pb-Pb ages of about 1,460-1,430 Ma have been determined using garnets
from granulite samples of the Epembe Unit (Seth et al., 2001), whereas the authors obtained
conventional and SHRIMP U-Pb ages of 1,810-1,635 Ma and 1,520-1,510 Ma for zircon cores
and rims, respectively. Unpublished U-Pb zircon age determinations from granitic gneiss and
amphibolite samples of the EC (Burger, in Menge, 1996) yield slightly younger ages of 1,470
Ma and 1,394 Ma, respectively. These ages are in good accordance with Pb-Pb garnet ages
obtained by Seth et al. (in prep.; 1,390-1,330 Ma) for samples of the amphibolite-facies Orue
Unit.
In Angola, the Pre-Eburnian (late Archean to early Palaeoproterozoic) basement rocks
were partly migmatized during the Eburnian orogeny (~ 2,200-1,800 Ma; Torquato & Oliviera,
1977) and intruded by large volumes of syntectonic granitoids (2,191 ± 60 Ma: Torquato et al.,
1979; 1,847 ± 62 Ma: Carvalho et al., 1979). Subsequently, these rocks were by themselves
transsected by numerous syn- to post-tectonic granitoids with ages ranging between ~ 1,800-
1 Introduction
23
1,600 Ma (e.g. Torquato & Oliviera, 1977; Carvalho et al., 1979, 2000; Torquato & Carvalho,
1992).
Basement rocks
Angola
Namibia
(Epupa Complex)
Kunene Intrusive Complex
and associated felsic rocks
Angola
Lithostratigraphical unit
Lithology
Age (Ma)
Regional granite
granite, granodiorite, tonalite
Bale group
Quibala-type granite
metasediments (conglomerate,
siltstone, arkose, sandstone,
greywacke, shale, quartzite)
granite, granodiorite
2,191 ± 60 (Rb-Sr 5 w.r.i.,
Torquato et al., 1979)
2,162 ± 99 (Rb-Sr 5 w.r.i.,
Torquato & Oliviera, 1977)
Eburnian gneiss
porphyroblastic gneiss
Leucocratic granite
granite
Caraculo granitoid
Bibala granitoid
Ruacana augengneiss
granitic gneiss
EC (undifferentiated)
granitic gneiss
EC (undifferentiated)
amphibolite
Epembe Unit
granulite-facies ortho- and
paragneisses
Orue unit
amphibolite-facies ortho- and
paragneisses
Kunene Basic Complex
troctolite
troctolite
Namibia
Mangerite dyke
mangerite
Red granite
granite
Kunene Intrusive Complex
leucogabbronorite
Swartbooisdrif Syenite
Qtz-syenite
syenite
Alkaline suite
Namibia
Epembe nepheline syenite
nepheline syenite
Epembe carbonatite
sövite
Swartbooisdrif carbonatitic breccia
carbonatitic breccia
Swartbooisdrif ferrocarbonatite
ferrocarbonatite veins
1,847 ± 62 (Rb-Sr 5 w.r.i.,
Carvalho et al., 1979)
1,790 ± 32 (Rb-Sr 6 w.r.i.,
1,763 ± 21 (Rb-Sr 8 w.r.i.,
1,686 ± 69 (Rb-Sr 1 w.r.i.,
Torquato & Carvalho, 1992)
1,596 ± 86 (Rb-Sr 4 w.r.i.,
Torquato & Oliviera, 1977)
1,795 +33/–29 (U-Pb zircon,
Tegtmeyer & Kröner, 1985)
~1,470 (U-Pb zircon, Burger, in
Menge, 1996)
~1,394 (U-Pb zircon, Burger, in
Menge, 1996)
~1,460-1,430 (Pb-Pb garnet, Seth
et al., 2001)
~1,520-1,510 (U-Pb zircon rims,
Seth et al., subm.)
~1,810-1,635 (U-Pb zircon cores,
Seth et al., subm.)
~1,390-1,330 (Pb-Pb garnet, Seth
et al., in prep.)
1,347 ± 13 (U-Pb zircon, Mayer et
al., 2000)
1,030 – 1,150 (internal Sm-Nd
orthopyroxene-plagioclase
isochron, Mayer et al., 2000)
1,370 ± 4 (U-Pb zircon, Mayer et
al., 2000)
~1,400-1,300 (Rb-Sr w.r.i.,
1,385 ± 25 (U-Pb zircon, Drüppel
et al., 2000)
1,376 (U-Pb zircon, Littmann et
al., in prep.)
1,345 (U-Pb zircon, Littmann et
al., in prep.)
1,216 ± 2.4 & 1,213 ± 2.5 (U-Pb
zircon, Littmann et al., 2000)
1,204 (U-Pb titanite, Littmann et
al., in prep.)
1,116 (U-Pb pyrochlore, Littmann
et al., in prep.)
1,138 (U-Pb pyrochlore, Littmann
et al., in prep.)
Table 1: Available geochronological data for Palaeo- to Mesoproterozoic rocks of north-western Namibia and
south-western Angola.
24
1 Introduction
A minimum age of the KIC has been provided by a Rb-Sr date of 1,190 ± 90 Ma using
muscovite from a pegmatite dissecting the anorthosite massif (Simpson & Otto, 1960;
recalculated by Cahen & Snelling, 1966). Recently, an internal biotite-plagioclase whole-rock
Rb-Sr isochron date of 1,347 ± 13 Ma has been determined for one anorthosite sample of the
Angolan part of the complex (Mayer et al., 2000). This age has been constrained by the same
authors by an almost concordant U-Pb age of 1,370 ± 4 Ma using zircons from a cogenetic
mangerite vein. These ages are in good accordance with an almost concordant U-Pb zircon age
of 1,385 ± 25 Ma for a dark anorthosite sample of the Namibian part of the KIC (Drüppel et al.,
2000). In contrast, two conspicuous internal plagioclase-orthopyroxene whole-rock Sm-Nd
isochrons of 1.03 Ga and 1.15 Ga have been obtained for anorthosite samples of the northern
Angolan part of the KIC by Mayer et al. (2000), which the authors interpret to reflect cooling
episodes or re-crystallization after the emplacement of the anorthosite body. With Rb-Sr wholerock ages of 1,350 ± 50 Ma (Carvalho et al., 1987) the granite belt (Red granite), that separates
the northern and southern part of the KIC, is coeval with the anorthositic rocks of the KIC.
With U-Pb zircon ages of ~1,380-1,340 Ma (Littmann et al., in prep.) the syenite dykes at
Swartbooisdrif, although post-dating the emplacement of the main anorthosite body, seem to
belong to the same magmatic event. In contrast, the ankeritic carbonatites of the Swartbooisdrif
area are distinctly younger as is evidenced by U-Pb pyrochlore ages of ~1,140-1,120 Ma
(Littmann et al., in prep.).
The nepheline syenite intrusions, exposed within the basement rocks in the Epembe area,
yielded U-Pb zircon ages of 1,216 ± 2.4 and 1,213 ± 2.5 Ma whereas an U-Pb zircon age of
1,204 Ma has been constrained for the E-W trending sövite dykes of the same area (Littmann et
al., 2000, in prep.). These age determinations rather suggest an independent formation of the
Epembe and Swartbooisdrif suites.
1.4 Previous work on the magmatic rocks of NW Namibia and SW Angola
Based on petrologic and geochemical investigations, the Kunene Intrusive Complex has
been interpreted as a massif-type anorthosite complex (Beetz, 1933; Simpson, 1970; Simpson &
Otto, 1960; Vermaak, 1981), as a layered intrusion like the Great Dyke of Zimbabwe (Stone &
Brown, 1958) and the Bushveld (Silva, 1990, 1992), as a transitional type between massif-type
and layered intrusion like the Michikamau intrusion (Carvalho & Alves, 1990; Menge, 1998) or
1 Introduction
25
as a composite massif-type anorthosite body (Ashwal & Twist, 1994; Morais et al., 1998; Slejko
et al., 2002). In addition, the petrogenesis of the white anorthosite has been interpreted as a
metasomatic transformation of the gneissic country rocks of the Epupa Complex (Köstlin, 1967,
1974). According to Köstlin (1974), the dark anorthosites intruded this white anorthosite massif
along a series of concentric faults. In contrast to Menge (1998) and Köstlin (1974) recent
detailed field and analytical studies of the Namibian part of the KIC (Drüppel, 1999) have shown
that both the white and the dark anorthosite suite are typical Proterozoic massif-type
anorthosites, the white anorthosite suite representing an older intrusion, crosscut by thick dykes
of cogenetic dark anorthosite along E-W trending faults. This interpretation is mainly supported
by (1) the excellent preservation of igneous mineralogy and textures in both anorthosite suites,
(2) sharp intrusion contacts between the two anorthosite suites, (3) inclusions of angular clasts of
white anorthosite up to 12 m across in the dark anorthosite and (4) high LREE abundances and
xenocrysts of andradite-rich garnet in white anorthosite indicate a high degree of crustal
contamination of this rock suite when compared to the dark anorthosite, consequently pointing to
a different evolution of the parental melts of the two suites (Drüppel & Okrusch, 2000).
In contrast to the KIC, the Epembe and Swartbooisdrif rock suites are not widely known
in the international literature. The sodalite occurrences were discovered by geologists of the
Bethlehem Corporation in the early 60´s. Simpson & Otto (1960) were the first to describe the
sodalite-bearing lithologies of Swartbooisdrif. The authors describe fine-grained intergrowths of
calcite, albite, sodalite and cancrinite with deep blue sodalite aggregates up to 20 cm in length.
Unfortunately they misinterpreted the magmatic carbonatite-fenite mixtures as sedimentary
deposits. A brief outline of the Swartbooisdrif carbonatite dykes is given in Martin (1965), who
first recognized the intrusive nature of the carbonatites. He emphasised the fan-like distribution
of the carbonatite dykes and the occurrence of conspicuous sodalite-cancrinite lenses within the
carbonatites. A more detailed investigation, including a petrographic description, was given by
Toerin (in Verwoerd, 1967). The author suggests, that both sodalite and cancrinite are of
metasomatic origin, with the sodalite replacing nepheline of nepheline syenite fragments and the
cancrinite replacing plagioclase of anorthositic xenoliths. The alkaline intrusive suite exposed in
the Epembe area has been investigated in some detail by Ferguson et al. (1975). Their study
concentrated on fenitic aureoles surrounding the nepheline syenite and carbonatite intrusions of
Epembe, which, in clear contrast to the Swartbooisdrif carbonatite dykes, lack the extensive
formation of sodalite. The latest review of the Swartbooisdrif sodalite occurrences is given by
Menge (1986, 1996), who found the ankerite-rich carbonatite dykes to intrude older and heavily
sheared syenites. He recognized the characteristic magmatic foliation of the carbonatite dykes
26
1 Introduction
with alternating layers dominated by ankerite, sodalite, analcite, cancrinite, albite or magnetite,
with the sodalite being formed by the interaction of carbonatitic fluids and the bordering
lithologies.
1.5 Methods of investigation
Sampling of the anorthositic rock units concentrated on marginal zones of the KIC since the
rugged topography of the Namibian KIC made a detailed investigation of its central parts
without helicopter aid impossible. Mapping was conducted on topographic sheets (1:50,000)
with the help of aerial photographs, landsat images and a Garmin Global Positioning System
(GPS).
Detailed mapping (1 : 5,000) was carried out on the sodalite mining area of Namibia Blue
Sodalite Ltd. (co.) in order to evaluate the complicate field relationships between the anorthosites
of the KIC and the syenites and ferrocarbonatites of the Swartbooisdrif area. Streets, dykes and
geological borders were doubly measured with the help of the Garmin GPS. The general course
of several larger carbonatite dykes was cross-checked with aerial photographs and landsat
images.
1.6 Aims of the study
Objectives of this study are the anorthositic rocks of the Kunene Intrusive Complex and the
spatially associated felsic suite as well as the younger carbonatites of the Swartbooisdrif area,
including sodalite-rich lithologies. The present study tries to give a comprehensive insight into
the Mesoproterozoic evolution of NW Namibia, with respect to the genesis and tectonothermal
evolution of these rock suites. The study has been subdivided into three sub-units in order to
solve certain special problems concerning the evolution of each individual rock unit:
The first part of this study outlines the complex field relationships between anorthosites, syenites
and carbonatites.
•
What are the structural controls for the emplacement of the anorthosites, syenites and
carbonatites?
•
Where does the sodalite exactly occur?
1 Introduction
•
27
Are there other rock units, which are possibly involved in the fenitizing processes?
The second part of this study is concerning with the petrogenesis of the anorthositic and the
felsic suite.
•
What are the source rocks of the anorthositic and the felsic rocks?
•
What are the pressure and temperature conditions during their emplacement,
crystallization and subsolidus re-equilibration?
•
Are the two suites consanguineous or just coeval?
•
Which processes led to their intrusion?
Part three of this study elucidates the magmatic and interrelated metasomatic processes of the
Swartbooisdrif carbonatites and related rock units.
•
What are the source rocks of the carbonatites?
•
Of what kind are the relationships between the carbonatitic breccia and the late
ferrocarbonatite veins?
•
Under which pressure-temperature conditions did the carbonatites crystallize?
•
Which processes led to the fenitization of anorthosites and syenites and the formation of
sodalite?
•
What is the nature and origin of the fenitizing solutions? Do they display an evolutionary
history?
•
What are the pressure-temperature conditions of the several mineral-forming,
metasomatic processes?
The only work on the relation between the anorthosites of the KIC, the associated felsic rocks
and the Swartbooisdrif sodalite deposits has been published by Menge in 1996, which however is
mostly based on investigations performed 20 to 30 years ago. Since knowledge in structural
geology, petrology and geochemistry tremendously increased during the last decades, a reexamination of the rock units using modern tools is most desirable.
28
2 Physiography of the area mapped
29
PART I: The geology of the sodalite mining area at Swartbooisdrif
2.1 Location of the area mapped
In order to evaluate the complicated relationships between anorthosite, syenite and ankerite
carbonatites, a locally restricted sub-area of the study area has been mapped in detail (scale: 1 :
5,000; Plate 1: geological map). This area is located about 5 km south-west of the NamibianAngolan border, marked by the river Kunene, and 6.8 km north-west of the settlement
Swartbooisdrif (Fig. 1.3). It comprises an area of about 12 km2, extending across longitudes of E
13°46.000´ to E 13°48.000´ and latitudes of S 17°20.000´ to S 17°21.520ánd is covered by a
map section of the topographic map 1:50,000 Swartbooisdrif (sheet 1713BD). The mapped
region is part of the mining area of Namibia Blue Sodalite Ltd. Company. It is transsected by the
ephemeral stream Ondoto and bordered by the NE-striking ridges of the Zebra Mountains in the
north-west.
2.2 Geographic and geomorphologic overview
The topography of the study area is closely related to the relative age and mineralogical
composition of the different lithologies and, to a certain extent, to regional tectonic structures.
The north-western part of the area mapped is dominated by ENE striking ridges with a strong
positive relief, consisting of weakly altered and dark coloured anorthositic rocks. These rugged
and steep mountains may rise to altitudes of about 940 m above sea-level and 150 m above the
bordering plain. Physical weathering of the anorthositic rocks led to the in-situ formation of
spheroidal boulders with diameters up to 10 m that cover the flanks of the ridges. Channelled
eluviation of soil along a radial drainage system is responsible for plant life localised as belts,
20-50 m apart. The alternation of the uncovered and dark-coloured boulder belts and vegetation
gives the ridges a striped appearance (Fig. 2.1a) which caused their regional name Zebra
Mountains.
In clear contrast, the undulating plain south-east of the ridges has an average altitude of 790
m above sea-level with the topography changing about 30-50 m in height over short distances. It
is underlain by white anorthosite as the dominant rock type of the major part of the study area.
The pervasively altered and strongly weathered white anorthosite is transsected by numerous SEand subordinate NE-trending tension-fissures, that were in part filled by younger syenite and
30
carbonatite dykes. The topography of the low-lying sub-area is mainly characterized by an
alternation of NW-striking and parallel running valleys and ridges that were most probably
formed as a result of the faster denudation of the tectonised and weathered white anorthosite
compared to the numerous younger syenite and carbonatite intrusions (Fig. 2.1b).
Fig. 2.1: a) Photography of the weathering resistant dark anorthosite ridges of the Zebra Mountains in the
background of an undulating plain, underlain by white anorthosite (Locality: S 17°20.417´, E 13°47.000´; NWdirection). b) Overview photography of the study area (Locality: S 17°20.075´, E 13°46.016´, S-direction), with the
two major dykes of the sodalite-rich carbonatitic breccia (dyke A and B) marked.
31
The courses of the river valleys of the ephemeral stream Ondoto and the numerous smaller
blind creeks are mainly controlled by the NE- and SE-trending fault system. Consequently, the
deeply eroded river bed of the Ondoto shows the lowest topographic altitudes of the study area,
i.e. 743 to 752 m above sea level. The drainage of the Ondoto is towards the Kunene River
which is the only perennial river in NW Namibia. The Kunene spruits in Angola, follows the
border of Angola and Namibia over a long distance and flows into the Atlantic. All blind creeks
of the study area as well as the radial drainage system of the Zebra Mountains empty into the
Ondoto.
The climate in NW Namibia is hot and subtropical. Rain falls mainly between December and
April with an average annual rainfall of 300 mm (Swartbooisdrif). The daily temperaturemaximum in the summer is 45°C, in the winter frost is common. As a consequence of these
climatic conditions and the dominance of physical weathering soil formation is rare.
The vegetation exhibits seasonal and regional diversity. During the dry winter months, the
barren vegetation is dominated by small bushes and trees (i.e. mopane: Colophospermum
Mopane and Copaifera Mopane, paper bark tree: Comuniphera, wild fig tree: Ficus Sycomorus,
and baobab: Adansonia Digitata), whereas the area is carpeted by several species of bushman
grass (Aristida Uniplumis) in the summer months.
2.3 Exploration history
The sodalite occurrences were discovered by geologists of the Bethlehem Corporation in the
early 60´s. The exploration of the sodalite as a semiprecious stone was started in 1969 by Bantu
Mining Corporation limited, Pretoria, South Africa. Between 1964 and 1979, about 700 t of
sodalite-rich rocks were mined, the pits started on sodalite outcrops subsequently developed into
small pits and opencastings. The use of ordinary TNT caused rupturing of the rocks and thus
gave rise to a long-term damage to the sodalite-bearing lithologies. In 1979 all mining activities
came to a standstill as a result of military operations in this area related to the Namibian-Angolan
civil war. In 1985 the production and export of sodalite-bearing lithotypes as blocks of
ornamental stone (1m x 1m x 2m), the Namibia Blue, was started. The mining area is now in
possession of the Namibia Blue Sodalite Ltd. Company.
32
3.1 Geological overview
3 Geology of the area mapped
The study area is located near the south-eastern margin of the KIC. The dominant rock
type in the central and southern part of the area mapped is a massive bright coloured anorthosite,
the white anorthosite (heavily altered anorthosite and leucogabbronorite [A,w]), with a general
white, green or violet appearance. In contrast, the north-western part is dominated by the steep
ridges of the Zebra Mountains that are mainly composed of less altered anorthosites and
leucogabbronorites [A,m] and leucotroctolites [T,m] of the dark anorthosite suite [A,d:
comprising A,m and T,m]. Smaller, subsequently emplaced intrusions of dark anorthosite within
the white anorthosite predominantly occur in central and eastern parts of the field area. In the
vicinities of the river bed of the Ondoto, the dark anorthosite is strongly brecciated and
weathered. Since it was impossible to assign the tectonised rocks to one particular anorthosite
unit of the dark anorthosite suite, the brecciated anorthosite [A,b] was classified as an
independent unit.
A conspicuous feature of the north-eastern part of the study area are elliptic
accumulations of ilmenite-magnetite ore, which are mainly exposed within the white anorthosite.
The general absence of exposed contacts between the weathered boulders of ilmenite-magnetite
ore [IM] and the white anorthosite makes it difficult to unravel their chronological succession,
yet the ores are suggested to be genetically and temporally linked to the anorthosite
emplacement.
Fractures and near-extensional shear zones in the anorthositic rocks of the KIC were
intruded by numerous SE and subordinate NE to ENE trending syenite and carbonatite dykes and
veins. In a study area of about 12 km2 a total of 288 dykes as well as numerous small veinlets of
strongly varying lithology were mapped. Individual dykes may be interrupted over short
distances and frequently change their strike and dip direction on a local scale.
The 82 felsic dykes are mainly represented by reddish to brownish, strongly altered Kfeldspar-rich syenite [S,a] which forms NE and SE striking dykes, crosscutting the anorthosite
complex. In shear zones intruded by both, syenites and carbonatites, the syenite is heavily
strained and exhibits a pink to brownish colour. The colour-change of the syenites may be
attributed to metasomatic exchange reactions between carbonatitic fluids and the syenite,
33
forming a fenitized syenite [S,f]. At one locality, a pegmatoidal nepheline syenite [S,ne] dyke,
crosscutting the syenite intrusions, was observed in direct contact to a carbonatite dyke.
A major swarm of carbonatite dykes and veins occur in the area, dissecting the syenite
dykes and the anorthosite body. The carbonatite dykes have variable trends with a predominant
SE, and a subordinate NE direction. Two main periods of carbonatite emplacement have been
recognized:
(1) The 187 dykes of the predominant carbonatite bodies, the carbonatitic breccia, are up
to 80 m in width and frequently contain variable amounts of angular to sub-rounded fragments of
fenitized wallrock anorthosite and syenite (70-95 vol.%), set in an ankeritic carbonatite matrix.
Most of the xenoliths are pervasively fenitized and partially broken to small particles making up
an interfragmental groundmass. A common feature of the carbonatitic breccia is a banded or
streaky appearance due to an alternation of ankerite-, magnetite- and silicate-rich layers oriented
sub-parallel to the dyke walls (layered carbonatitic breccia [Cb,l]). Flow-banding and
impersistent magmatic folding on a millimetre to meter scale are common structures of this rock
type. In places, where the grain size of the rocks decreases to very fine-grained, following syn- to
post-emplacement shearing, the rocks exhibit a massive texture and a homogeneous grey colour
(massive carbonatitic breccia [Cb,m]). Ankerite, albite and magnetite are the main minerals
present in the carbonatitic breccia, although calcite, dolomite, cancrinite, biotite, muscovite,
apatite, sulphides and sodalite are locally observed in significant concentrations. Variable
amounts of sodalite occur as conspicuous deep blue lenses, layers and breccias in several of the
larger dykes of the carbonatitic breccia (sodalite-rich carbonatitic breccia [Cb,so]), but are also
abundant as metasomatic aureoles up to 1 m in width within the bordering dark anorthosite. Dark
coloured K-feldspar-biotite-ilmenite-calcite rocks, strongly resembling lamprophyres but yet
displaying conflict relationships to the carbonatitic breccia, are present in minor proportions.
(2) Both, the main carbonatite body and the K-feldspar-biotite-ilmenite-calcite rocks are
transsected by and intermingled with small veins and stringers of a second generation of almost
silicate-free ferrocarbonatite [FV] veins and stringers, which may be present in strongly varying
proportions. In contrast to the carbonatitic breccia, no effects of metasomatic alteration of the
bordering rocks are visible on a macroscopic scale.
Sparse SE trending dolerite [Dol] dykes up to 70 m in width, which are most abundant
throughout the eastern part of the investigated field area, cut both the syenites and the
carbonatites. In addition, one andesite dyke [An] and one dyke of a carbonate-rich trachyte [Tr]
34
have been mapped. Due to their limited aerial extent, it was impossible to constrain clear
temporal and structural relationships with the carbonatite, syenite and dolerite dykes.
The magmatic lithologies are overlain by different sedimentary rock units. Anorthositic
talus deposits cover the steep ridges of the Zebra Mountains in the north-western part of the
study area. The riverside of the ephemeral stream Ondoto is flanked by unconsolidated quartzite
pebble accumulations [Qtz,p], which most probably represent variably eroded relics of former
fluvial terraces. The youngest fluvial terrace generation, carbonate-cemented immature
conglomerates [Cgl], border the Ondoto riverbed. The river floor of the Ondoto is filled by
barren sand [All] intercalated with angular anorthosite and carbonatite fragments.
Mining waste piles in the vicinity of abandoned and active opencastings and the
settlement Orotumba testify to the continuous pressure-production history.
3.2 Rock exposure
Due to the vegetation cover, many of the carbonatite outcrops are not accessed easy. In addition,
the strong oxidation of ankerite under atmospheric conditions causes homogeneous brownish
weathering surfaces of the carbonatitic breccia and thus makes it difficult to identify its
mineralogy and texture. Good exposure of carbonatites and syenites and their contact
relationships is confined to the numerous disused sodalite quarries formed by previous mining
operations and the deeply eroded river valley of the Ondoto river. Excellent exposure is provided
by the presently active sodalite quarry.
Since the KIC mountain ridges are generally covered by vegetation and anorthosite debris
whereas the plains underlain by white anorthosite are covered by soil, mapping and sampling of
the anorthositic rock types of area mapped concentrated on sodalite-quarries, the Ondoto river
valley and scattered leucotroctolite outcrops occurring in the north-western part of the study
area.
3.3 Field observations
35
3.3.1 The white anorthosite suite [A,w]
The study area is dominated by the white anorthosite that represents the oldest magmatic rock
unit. In outcrops, the strongly tectonised white anorthosite displays a massive appearance, due to
a homogeneous structure without preferred orientation of the coarse-grained plagioclase crystals
(Fig. 3.1a). Accumulations of Fe-Ti-Oxides, associated with biotite and/or amphibole are rare.
Fig. 3.1: a) Outcrop of massive white anorthosite, exposed in the river valley of the ephemeral stream Ondoto
(Locality: S 17°20.437´, E 13°46.860´). b) An angular clast of white anorthosite is enclosed by the grey coloured
dark anorthosite (Locality: S 17°20.019´, E 13°46.529´). c) In the carbonatitic breccia, clasts of the white anorthosite
are wrapped by brownish ankerite-rich (Locality: S 17°21.045´, E 13°47.078´) or d) grey to blue silicate-rich layers
(Sodalite opencasting 7, locality: S 17°20.850´, E 13°46.824´).
The hololeucocratic white anorthosite is characterized by a whitish to greenish or violet
colour due to a pervasive sericitization and saussuritization of plagioclase and an alteration of
Fe-Mg silicate phases. If exposed, intrusion contacts between the white anorthosite and the dark
anorthosite are always sharp; inclusions of angular clasts of white anorthosite up to 12 m across
in the younger dark anorthosite are a common feature in the investigated field area (Fig. 3.1b). In
36
addition, rotated and elongated fragments of the white anorthosite of up to 3 m across occur
within the carbonatitic breccia (Fig. 3.1c, d).
Due to its strong alteration and brecciation the white anorthosite is generally weathered
and deeply eroded, with the undulating erosion plains being covered by soil (Fig. 2.1a).
Exposures of solid bedrock are rare, except in the Ondoto river valley and several disused
sodalite quarries.
3.3.2 The dark anorthosite suite
3.3.2.1 Anorthosites and leucogabbronorites [A,m]
Irregular bounded anorthosite and leucogabbronorite (including leucogabbro and leuconorite)
mounds with diameters of up to 300 m occur within the white anorthosite. These rocks dominate
in the western part of the study area but may also occur in the centre and eastern regions. Due to
their grey to greenish-grey colour and their resistance against alteration and weathering these
rocks can be well distinguished from the white anorthosite on a macroscopic scale.
The anorthosites and leucogabbronorites are characterized by a massive texture with
subhedral, coarse-grained plagioclase crystals of up to 26 cm in length (Fig. 3.2 a). Cumulations
of Fe-Mg-silicates and Fe-Ti oxides of up to 0.5 m in diameter are relatively widespread on a
local scale (Fig. 3.2 b). Leucogabbronorites can be distinguished from the anorthosites by their
higher modal amounts of Fe-Mg silicates, i.e. clinopyroxene, orthopyroxene, biotite and
amphibole. It has to be mentioned, however, that the boundaries between anorthosite and
leucogabbronorite are generally transitional, thus the two rock units could not be distinguished as
independent lithologies in the field.
Both rock types generally display sharp contacts against the older white anorthosite and
the younger leucotroctolite. Commonly, dykes and small plugs of the leucotroctolites crosscut
the older anorthosites and leucogabbronorites (Fig. 3.2 c). In the Ondoto river valley, the rocks
may display a low angle banking structure which is most probably related to joint tectonics
rather than representing magmatic layering as was proposed by Menge (1986).
Attaining up to 65-98 vol.%, plagioclase is the major constituent of these rock types.
Minor phases are amphibole, clinopyroxene, orthopyroxene, biotite and Fe-Ti oxides with
strongly varying amounts. When approaching the contact to dykes of the carbonatitic breccia,
plagioclase is successively replaced by sodalite along fractures leading to a colour change of the
37
rocks from grey into purplish blue. In addition, accessory sulphides may fill small cracks in the
anorthositic rocks. At the direct contact to the carbonatite dykes, the anorthositic rocks display a
strong foliation oriented sub-parallel to the intrusion interfaces. In these zones, elongated
plagioclase crystals are almost completely replaced by sodalite (Fig. 3.2 d), the anhydrous FeMg silicates are altered to biotite and the amount of carbonate minerals increases drastically.
Fig. 3.2: a) Outcrop of the grey coloured dark anorthosite in the Ondoto riverbed. Note the homogeneous texture of
the anorthosite, containing a plagioclase megacrysts of 18 cm length (Locality: S 17°20.019´, E 13°46.529´). b)
Accumulations of amphibole(brownish) in a transitional anorthosite/leucogabbronorite of the dark anorthosite suite,
exposed at the south-western margin of the KIC (Locality: S 17°08.405´, E 13°14.181´). c) Narrow leucotroctolite
dykes (T,m) cut older leucogabbronorites (A,m) of the dark anorthosite suite (Locality: S 17°20.286´, E
13°46.224´). d) The leucogabbronorite (left) in direct contact to the carbonatitic breccia (right) displays a weak
foliation marked by ankerite- (brown), biotite- (black), sodalite- (blue) and plagioclase-rich (white) schlieren
(Sodalite opencasting 8, locality: S 17°20.470´, E 13°47.000´).
3.3.2.2 Leucotroctolites [T,m]
Leucotroctolite [T,m], including leucotroctolite-anorthosite transitions, occurs near the western
boundary of the study area as 19 independent rounded to oval shaped intrusions or narrow dykes
38
within both, the white anorthosite and the anorthosites/leucogabbronorites of the dark anorthosite
suite. The biggest of these intrusions has a diameter of approximately 180 m. In addition,
leucotroctolite occurs at the north-western boundary of the study area where it forms spectacular
bolder belts covering the steep ridges of the Zebra Mountains (Fig. 3.3 a).
In fresh exposures leucotroctolites are characterized by dark grey to black colour. In clear
contrast, the rocks display sacklike structures and brown coloured alteration rims in outcrops
exposed to weathering. The leucotroctolites are commonly medium-grained with a homogeneous
texture lacking a preferred orientation of the plagioclase crystals (Fig. 3.3 b). A conspicuous
feature of the leucotroctolites is the strong iridescence of their plagioclase crystals, which makes
these rocks most suitable for an exploitation as dimension stones.
The mineralogy of the leucotroctolites is much more restricted than that of the
anorthosites and leucogabbronorites. Plagioclase (89-98 vol.%) is the dominant mineral
accompanied by minor olivine (up to 8 vol.%). Accessory pyroxene, amphibole, biotite and FeTi oxides may occur even though cumulate textures, as common in the leucogabbronorite, are
lacking.
3.3.2.3 Brecciated anorthosites [A,b]
In the vicinity of the Ondoto river as well as along the numerous fault-related streamlets the
anorthositic rocks are strongly tectonised. The diameter of the angular fragments rarely reaches
1.5 cm. The shatter zones are mainly composed of a porous plagioclase framework with the
altered Fe-Mg silicates and Fe-Ti oxides being weathered and washed out (Fig. 3.3 c).
Contacts between the brecciated anorthosite and its undeformed counterparts are always
sharp. Therefore, the brecciated anorthosites can not be assigned to one particular anorthosite
unit and were thus mapped as an independent unit. The fact that narrow syenite dykes within the
brecciated anorthosite are not affected by this deformation, strongly suggests that the brecciation
of the anorthosites and the formation of the branched fault system predates the syenite
emplacement.
3.3.3 Ilmenite-magnetite ore [IM]
Lensoidally bounded accumulations of weathered ilmenite-magnetite ore [IM] of up to 125 m in
diameter particularly occur within the white anorthosite in the north-eastern part of the study
39
area but may also be present within leucogabbronorites and anorthosites of the dark anorthosite
suite. If exposed, contacts between the black coloured Fe-Ti ore and the host anorthosite are
generally transitional. However, outcrops of solid bedrock are generally rare in the study area,
since the ilmenite-magnetite ore is mostly weathered to rounded pebbles which are up to 8 cm in
diameter (Fig. 3.3d).
Fig. 3.3: a) Leucotroctolites exposed in the Zebra Mountain massif form boulder belts, that cover the ridges in
zones of up to 20 m apart (Locality: S 17°20.126´, E 13°46.524´, W-direction). b) The commonly medium-grained,
dark grey to black leucotroctolites may contain plagioclase crystals of up to 8 cm in length (Locality: S 17°26.830´,
E 13°37.995´). c) In the vicinity of a dyke of the carbonatitic breccia, the anorthositic rocks are brecciated.
Plagioclase forms a porous framework, whereas the intercumulus Fe-Mg silicates and Fe-Ti oxides have been
washed out (Locality: S 17°19.904´, E 13°46.601´). d) Deposits of the ilmenite-magnetite ore are commonly
exposed as accumulations of dark brown, rounded pebbles (Locality: S 17°20.126´, E 13°46.524´).
Ore minerals mainly consist of magnetite and ilmenite (> 90 vol.%) with the magnetite
displaying macroscopically visible ilmenite exsolution lamellae. Interstices between the Fe-Ti
oxide grains are filled by sulphides and relics of heavily altered Fe-Mg silicates. Cavities in the
weathering surfaces of the ore boulders suggest that superficial Fe-Mg silicates were leached out.
40
The fact, that the exact 3-dimensional shape of the ore bodies is yet unknown and
exposed wallrock contacts are rare, makes it difficult to unravel their temporal succession and
their genetic setting. Most likely, however, the generation of the ilmenite-magnetite ores is
genetically linked to the intrusion of the white anorthosite and the leucogabbronorite. The Fe-Ti
ores may have separated by either (1) mechanical segregation of crystallising Fe-Ti oxide
minerals following gravity settling or filter pressing or (2) oxide magma immiscibility followed
by crystallization of Fe-Ti oxides from the separated liquid phase. Based on field investigations
Morais et al. (1998) rather preferred the second model for similar ore bodies in the Angolan part
of the KIC. Similar Fe-Ti ores have been found to be associated with anorthosites world-wide
(e.g. Duchesne, 1999).
3.3.4 Syenites [S,a] and [S,f]
In the study area 82 major syenite dykes with a length of 10-650 m and a width of 15 cm to 15 m
were mapped, not including numerous narrow veins and apophyses extending into the
anorthositic rock units, which were too small to show them on the geological map. The syenites
may be subdivided into two units:
(1) The vast majority of the felsic intrusives occurs in the south-eastern area as reddish to
brownish, strongly altered and weathered K-feldspar-rich syenite to leucosyenite [S,a] which
forms NE and subordinately SE trending dykes, exposed up to 600 m in length and 10 cm to 15
m in width and crosscutting the anorthosite complex. In most cases these syenites are heavily
brecciated and/or sheared into small particles. Outcrops of solid bedrock are rare, commonly the
syenites form accumulations of angular fragments (Fig. 3.4 a, b). When exposed in reactivated
shear zones, the syenites are altered to epidote which may also be present as massive epidote
rocks replacing the syenite dykes as a whole.
The medium-grained syenites are characterized by a homogeneous texture (Fig. 3.4b)
without preferred orientation of the K-feldspar crystals (75-90 vol.%). Minor constituents are
amphibole, biotite and Fe-Ti oxides. Contact-metamorphic phenomena have never been
observed in the anorthositic rocks bordering the syenite dykes nor are chilled margins developed
in the syenite itself. However, narrow syenite dykes may display a flow zoning, with the biotite
crystal sizes increasing towards the dyke centre (Fig. 3.4c). It may thus be concluded that the
anorthosites still had elevated temperatures during the syenite emplacement. However, in some
of the dark anorthosite occurrences, syenitic fluids led to the contact-metasomatic formation of
disseminated sulphides.
41
Angular fragments of syenite are incorporated in the carbonatitic breccia, locally reaching
amounts of up to 40 vol.% (Fig. 3.4d). The brecciated and rotated xenoliths may approach
diameters of up to 4 m. Since these syenite clasts are too big to be transported over long
distances by the carbonatitic melt, the most successful model for the carbonatite emplacement is
a passive intrusion mechanism following fault-related brecciation of the syenite.
Fig. 3.4: a) Syenite commonly forms accumulations of weathered, angular syenite fragments (Locality: S
17°21.145´, E 13°47.772´). b) In rare cases, the reddish syenite is exposed in small outcrops of solid bedrock
(Locality: S 17°21.145´, E 13°47.772´). c) A narrow biotite-bearing syenite dyke (brownish) transsects an older
leucogabbronorite (grey) of the dark anorthosite suite. The crystal sizes of biotite successively increase towards the
centre of the syenite dyke (Locality: S 17°21.072´, E 13°46.495´). d) Rotated fragments of the heavily tectonised
syenite (pink) occur within the carbonatitic breccia, where they are wrapped by carbonate- (brown) and silicate-rich
(grey, blue) layers (Locality: S 17°21.944´, E 13°48.180´).
(2) In the predominantly SE trending shear zones intruded by both, syenite and
carbonatite, the syenites are heavily strained and mostly exhibit a pink colour (Fig. 3.5 a, b). The
syenite occurs within and at the borders of the dykes of the carbonatitic breccia with a width
ranging from < 3 m up to 15 m. The exposed contacts between syenite and carbonatite are
42
curvilinear (Fig. 3.5 a) but tectonic, since syenitic fragments in all stages of detachment can be
observed in the bordering carbonatite (Fig. 3.5 b).
In clear contrast to the K-feldspar-rich syenite dykes, the coarse-grained fenitized
syenites [S,f] are predominantly composed of subhedral albite crystals of up to 3 cm length,
already visible on a macroscopic scale. When approaching the contacts to the carbonatite dykes,
the fenitized syenite displays a weak foliation sub-parallel to the dyke walls. Joints in the
fenitized syenites are commonly filled by clusters of euhedral magnetite crystals of up to 3 cm in
length.
The colour-change of the syenites must be attributed to metasomatic exchange reactions
between Na-rich carbonatitic fluids and the K-feldspar-rich syenite forming albite-rich fenitized
syenites. It has to be mentioned, however, that the rarely exposed contacts between the syenite
and the fenitized syenite are always sharp, the latter dissecting the former. Thus it may be
concluded that the predominantly SE trending fenitized syenites are not only fenitized
equivalents of the syenite but may additionally represent a second syenite generation emplaced
in a younger, SE trending fault bundle.
Common features of both syenite dyke swarms are branching, dying out and curving of
the dykes as well as dyke segmenting into an en-echelon set of intrusions, with the width of the
dykes continuously decreasing from >10 m down to several tens of cm towards their
terminations.
3.3.5 Nepheline Syenite [S,ne]
At one locality, in the eastern part of the mapped area (Locality: S 17°21.100´, E 13°47.850´), a
nepheline syenite dyke, crosscutting a syenite intrusion, was observed in direct contact to a dyke
of the carbonatitic breccia. The dyke is exposed over just a short distance of about 6.5 m and a
width of 3 m to 30 cm.
At distance to the carbonatite dykes, the pink nepheline syenite is characterized by a
homogeneous texture with subhedral and medium-grained nepheline and K-feldspar being the
major constituents, accompanied by minor biotite and Fe-Ti oxides. When approaching the
carbonatite dykes, the texture of the nepheline syenite successively becomes inhomogeneous
with fine-grained K-feldspar-rich zones alternating with irregular pegmatoidal zones, mainly
composed of nepheline crystals of up to 5 cm in length (Fig. 3.5c). At the direct contact to the
43
carbonatite dykes the nepheline syenite exhibits a massive texture with euhedral nepheline
crystals of up to 15 cm across. In these zones, nepheline exhibits a pale blue tint resulting from
its partial replacement by sodalite (Fig. 3.5 d). The increase of the nepheline crystal size at the
contact to the carbonatite dyke and the replacement of nepheline by sodalite may be attributed to
the late influx of Na-rich carbonatitic fluids, but may also be due to the cogenetic formation of
the two rock types.
Fig. 3.5: a) From distance, the contacts between the pink coloured fenitized syenite and the carbonatitic breccia
(dark grey) look curvilinear, thus suggesting a cogenetic origin of the two rock units (Locality: S 17°20.970´, E
13°47.113´). b) However, in a more detailed view, syenite fragments (pink) in all stages of detachment can be
observed in the bordering carbonatitic breccia (black), testifying to the intrusive nature of the younger carbonatites
(Sodalite opencasting 7, locality: S 17°20.890´, E 13°46.774´). c) The nepheline syenite in contact to the
carbonatitic breccia displays an inhomogeneous texture, characterized by alternating K-feldspar- and nepheline-rich
schlieren (Locality: S 17°21.101´, E 13°47.854´). d) At the direct contact to the carbonatitic breccia, euhedral
nepheline of the nepheline syenite exhibits a blue tint, resulting from its partial replacement by metasomatically
formed sodalite (Locality: S 17°21.101´, E 13°47.854´).
44
3.3.6 Carbonatites and associated rock types
3.3.6.1 Carbonatitic breccia
A major swarm of 199 carbonatite dykes and veins occurs in the area, dissecting the syenite
dykes and the anorthosite body. The carbonatite dykes have variable trends with a predominant
SE, and a subordinate NE direction. Like the syenites, the dykes of all three carbonatite
subgroups display features like dying out, branching, curving and en-echelon segmentation. The
predominant carbonatite bodies, the carbonatitic breccia, are up to 80 m in width and frequently
contain variable amounts of angular to sub-rounded fragments of wallrock anorthosite and
syenite (~70-95 vol.%). The fragments range continuously from 8 m down to several millimetres
in diameter, set in an ankeritic carbonatite matrix. Most of the xenoliths are pervasively fenitized
and partially broken to small particles making up an interfragmental groundmass. Regarding
both their macroscopic appearance and their general mineralogy, three subtypes of the
carbonatitic breccia were established:
(1) The vast majority of the dykes of the carbonatitic breccia, i.e. 145 dykes, displays a
banded or streaky appearance due to an alternation of ankerite-, magnetite- and silicate-rich
layers oriented sub-parallel to the dyke walls. Flow-banding and impersistent magmatic folding
on a millimetre to meter scale are common structures of the layered carbonatitic breccia [CB,l].
Sodalite is a minor constituent or even absent.
(2) In 45 more medium to coarse-grained dykes of the carbonatitic breccia, sodalite is
concentrated in economic proportions of > 25 vol.%. Variable amounts of sodalite occur as
conspicuous deep blue lenses, layers and breccias within the larger carbonatite dykes, but are
also abundant as metasomatic aureoles up to 1 m in width within the bordering dark anorthosite.
These rocks are referred to as sodalite-rich carbonatitic breccia [CB,so].
(3) In clear contrast, 12 dykes of the carbonatitic breccia are characterized by a massive
appearance and a homogeneous grey colour with rare reddish feldspar-rich xenoliths. The grain
sizes of the massive carbonatitic breccia [CB,m] range from fine-grained to very fine-grained,
most probably following syn- to post-emplacement shearing. Sodalite has never been observed in
these rocks on a macroscopic scale.
3.3.6.1.1 Layered carbonatitic breccia [CB,l]
Dykes of the banded carbonatitic breccia are spread over the whole study area. They are exposed
over a length of 5-800 m with a varying thickness ranging between 20 cm and 18 m. The rocks
are characterized by an impersistent layering sub-parallel to the wallrock interface, marked by an
45
alternation of bands rich in, or wholly composed of albite (pink to light grey) and biotite (dark
brown to black), carbonates (brownish to yellowish) and oxides (black) whereas sodalite (blue),
cancrinite (light yellow) and sulphides are minor constituents (Fig. 3.6 a, e, f).
Fig. 3.6: a) Sub-parallel alternation of grey albite- and biotite-rich and brownish ankerite-rich layers of the layered
carbonatitic breccia, containing dispersed, sodalite-rich xenoliths (Locality: S 17°20.066´, E 13°46.524´). b) A dyke
of the layered carbonatitic breccia (dark brown), displaying impersistent lamination, transsects a leucogabbronorite
(grey), which, by itself, is intruded by a syenite dyke (pink) (Sodalite opencasting 7, locality: S 17°20.890´, E
13°46.774´). c) and d) The sub-parallel layering of the carbonatitic breccia, with the laminae wrapping rotated
anorthosite (light grey), sodalite (blue) and syenite fragments (pink), gives the rock a mylonite-like appearance
(Locality: S 17°20.147´, E 13°46.662´). e) Common features of the carbonatitic breccia are the displacement of
already completely crystallized laminae along impersistent brittle faults and f) discontinuous magmatic folding
(Sodalite opencasting 4, Locality: S 17°20.967´, E 13°47.722´).
46
Contacts between individual layers may be both sharp or transitional (Fig. 3.6 b). Rotated
angular fragments of altered anorthosite with sericitized plagioclase crystals (white to light grey;
Fig. 3.6 c) and syenite (reddish to light pink; Fig. 3.6 d), with diameters ranging between several
millimetres and 2.5 m, are commonly enveloped by ankerite-magnetite layers. Single laminae of
the banded carbonatite breccia may run sub-parallel over a certain distance, giving the rock a
mylonite-like appearance. In places, the laminae are displaced along small scaled brittle faults
which die out after short distances of 5-90 cm (Fig. 3.6 e). In addition, the dykes may display
discontinuous magmatic folding, with small scaled ductile shear zones being developed locally
(Fig. 3.6 f). The occurrence of carbonatite rafts, oriented discordant to the magmatic foliation of
the surrounding carbonatite matrix, supports evidence for repeated injections of carbonatite melt.
The fact that ankeritic laminae fill even the narrowest joints in the bordering anorthosite and
syenite prove that the carbonatite melt itself had a low viscosity, even though the textures
displayed by carbonatite-xenolith mixtures, as described above, closely resemble those of high
viscosity melts like rhyolites.
The numerous abandoned and active opencastings are exceptionally suitable for the
investigation of the textural features described above (Fig. 3.6 b, e, f). In addition, these textures
can be observed in outcrops exposed to weathering since the minerals contained in the
carbonate- and the silicate-rich layers display a strongly differing weathering behaviour (Fig. 3.6
a, c, d).
3.3.6.1.2 Sodalite-rich carbonatitic breccia [CB,so]
Economic concentrations of sodalite occur in several of the larger carbonatite dykes exposed
over a length of up to 2.8 km and a varying width of 10-90 m. Textural features of layered
sodalite-rich lithologies resemble those of the sodalite-poor subgroup (see above). In contrast to
the latter, sodalite-rich carbonatites may additionally contain massive ankerite-magnetite patches
with subordinate contamination by wallrock-material and/or massive fenite zones, that are
almost completely composed of albite and biotite. Within the carbonatite dykes, sodalite occurs
in three different textural positions: (1) As pale blue monomineralic layers of up to 1.2 m in
width in which sodalite replaces fine-grained, fragmented plagioclase (Fig. 3.7 a), (2) as deep
blue and partially translucent sodalite pseudomorphs after fragmented plagioclase-rich xenoliths
of up to 1.5 m in diameter enveloped by carbonate-rich laminae or massive albite-biotite fenite
zones (Fig. 3.7 b) and (3) as deep blue to green sodalite aggregates partially replacing
47
subrounded, feldspar-rich fragments dispersed in the massive ankerite-rich patches (Fig. 3.7 c, d;
Fig. 3.8 a, b).
Fig. 3.7: a) A monomineralic layer of deep blue sodalite is exposed within a dyke of the sodalite-rich carbonatitic
breccia (grey; Locality: S 17°19.904´, E 13°46.601´). b) Low-viscosity carbonatite melts (light to dark brown)
intruded a heavily brecciated anorthosite(grey)-syenite(pink) assemblage. Dark anorthosite fragments are partially
replaced by sodalite and wrapped by carbonate-rich laminae (Sodalite opencasting 7, locality: S 17°20.890´, E
13°46.774´). c) Sodalite-rich xenoliths (blue) are homogeneously dispersed in massive ankerite-rich patches (brown;
Locality: S 17°20.567´, E 13°47.605´). d) Same as (c) (Locality: S 17°20.902´, E 13°45.823´). e) Trails of magnetite
(black) mark the sodalite (blue) / carbonatite (yellowish) contacts (Sodalite opencasting 8, locality: S 17°20.417´, E
13°47.000´). f) A rim of medium-grained biotite flakes (black) is developed around sodalite fragments (blue),
enclosed by fine-grained albite-biotite zones (dark grey; Locality: S 17°20.902´, E 13°45.823´).
48
Trails of euhedral magnetite crystals mark the contacts between sodalite- and ankerite-
rich zones (Fig. 3.7 e), whereas medium to coarse-grained biotite flakes crystallize in the vicinity
of sodalite fragments, surrounded by fine-grained albite-biotite-rich zones (Fig. 3.7 f). The
sodalite distribution within the dykes of the carbonatitic breccia is quite heterogeneous. Even
within single dykes the amounts of sodalite commonly vary between 15 and 90 vol.%. In rare
cases, sodalite may also be absent over a certain distance, a fact that complicates the mine
layouts. It has to be mentioned, however, that sodalite is mostly difficult to identify in
carbonatite outcrops exposed to weathering, since the oxidation of ankerite causes the formation
of brownish alteration rims that may cover the whole rock. Sodalite concentrations of >90 vol.%
containing sodalite fragments of up to 0.5 m in diameter predominantly occur in regions, where
several large carbonatite dykes intersect at acute angles and the modal abundance of wallrock
xenoliths increases, thus evidencing that sodalite is not a magmatic crystallization product of
the carbonatites themselves but originates from the interaction of carbonatitic NaCl-rich fluids
with the incorporated feldspar-rich xenoliths. This interpretation is constrained by the occurrence
of sodalite in narrow metasomatic aureoles of up to 1 m in width within the anorthosites, that
surround large dykes of the carbonatitic breccia (Fig. 3.2 d). Heavily tectonised cross-points of
several large fractures intruded by large carbonatite dykes (geological map) seem to form
exceptionably suitable pathways for large-scale fluid-flow.
In the following, a short description of the general characteristics of the 9 major active
and abandoned opencastings, situated along the two major sodalite-rich dykes of the study area
(dyke A in the north-eastern part and dyke B in the central part of the study area; Fig. 2.1a) will
be given:
Sodalite opencasting 1 (dyke A, S 17°20.437´, E 13°47.423´)
In the abandoned opencasting (∅: 40 m x 100 m) white anorthosite is transsected by a more or
less E-W striking dyke of the sodalite-rich carbonatitic breccia dipping 30-35° to the S. The
weathered dyke is exposed over a width of ~10 m and mainly composed of massive and coarsegrained ankeritic layers, alternating with narrow and fine-grained grey laminae, oriented subparallel to the dyke walls. Numerous sub-rounded and feldspar-rich xenoliths, partially replaced
by sodalite, are scattered in the ankeritic zones giving the rock a speckled appearance.
Anorthositic wallrock of the foot wall are tectonised whereas anorthosites of the hanging wall
display a dyke-parallel foliation and a partial sodalitization of the elongated plagioclase crystals,
most probably related to upward moving carbonatitic fluids. Metasomatic changes of the
anorthositic mineralogy were observed in up to 1.5 m distance to the dyke contacts.
49
Sodalite opencasting 2a (dyke A, S 17°20.563´, E 13°47.605´ to S 17°20.599´, E 13°47.614´)
The disused sodalite digging (∅: 140 m x 50 m) is in its main parts filled by mining debris. Due
to the use of TNT during the early exploration activities, the outcropping carbonatite dyke is
heavily shattered. The SE striking carbonatite dyke dips with angles between 35 to 70° to the SW
and is dominated by brownish weathered, coarse-grained ankerite intermingled with large
amounts of sub-rounded sodalite-rich wallrock fragments (~ 75 vol.%) with diameters of up to 1
m. Towards the hanging wall contacts, the carbonatite displays a discontinuous narrow foliation
and asymmetric magmatic small scale folding. Anorthositic rocks in contact to the carbonatite
display a dyke-parallel foliation.
Sodalite opencasting 2b (dyke A, S 17°20.553´, E 13°47.685´)
The abandoned opencasting (∅: 20 m x 10 m) is an oval mine cavity in the centre of a 35-45° W
to NW dipping dyke generated by blasting operations. The opened dyke forms the eastern
extension of the carbonatite exposed in the sodalite opencastings (1) and (2). The outcropping
carbonatite displays a banded texture, with sodalite being a major constituent (average amount:
45 vol.%).
Sodalite opencasting 3 and 4 (dyke A, S 17°20.790´, E 13°47.689´ - S 17°20.967´, E
13°47.772´)
In the disused opencastings (∅: 10 m2) massive anorthosites and leucogabbronorites are
transsected by a SSE striking dyke of the carbonatitic breccia dipping with angles between 30°
and 50° to the SW. The 7-9.5 m thick dyke is characterized by an alternation of deep blue
sodalite-rich laminae (ca. 55 vol.%) and brownish carbonate-rich layers, oriented sub-parallel to
the intrusion contacts (Fig. 3.6 e; Fig. 3.20 c). The carbonatitic breccia exposed in sodalite
opencasting 3 additionally exhibits a weak lineation (Fig. 3.20 d). Impersistent magmatic folding
of the laminae dominates the centre parts of the dyke (Fig. 3.6 f; Fig. 3.20 a). In the bordering
anorthositic rock units, no effects of a metasomatic alteration have been observed. The dyke is
transsected by a younger generation of almost silicate-free ferrocarbonatite veins, that dissect the
magmatic foliation of the main carbonatite body.
Sodalite opencasting 5 (dyke B, S 17°21.492´, E 13°47.509´)
The small pit (5 m x 3 m) cuts a 8 m thick dyke of the sodalite-rich carbonatitic breccia that dips
with 30-52° to the NE and transsects the white anorthosite. In its central parts, the dyke is
composed of a massive ankerite-rich zone that bears conspicuous light green and partly
50
translucent sodalite fragments (Fig. 3.8 a, b). The colour of these sodalite fragments may change
from green to deep blue on a cm-scale (Fig. 3.8 a). Towards the dyke margins the massive
ankerite zone is bordered by folded silicate- and carbonate-rich laminae.
Sodalite opencasting 6 a (dyke B, S 17°21.045´, E 13°47.078´)
The abandoned opencasting, situated ca. 120 m south-east of a former military fort, extends over
an area of 15 m x 10 m. The mining operations concentrated on a 23 m thick sodalite-rich
carbonatite dyke dipping 35-45° to SW with sodalite-accumulations, enriched as layers. In the
strongly weathered carbonatite dyke anorthosite and fenite clasts have been observed, which are
surrounded by the foliated carbonatite matrix. A contact aureole up to 5 m thick is developed in
the neighbouring anorthosite, in which a contact parallel foliation is developed within the
anorthosite.
Fig. 3.8: Conspicuous light green coloured sodalite was observed at one locality (Sodalite opencasting 5, locality: S
17°21.492´, E 13°47.509´). a) The colour of sodalite may change from green to deep blue on a cm-scale b) Both the
green and the blue sodalite are concentrated in the central, ankerite-rich zones (brown) of a carbonatite dyke.
Sodalite opencasting 6 b (dyke B, S 17°21.006´, E 13°47.010´)
In this opencasting, the carbonatite dyke of the sodalite opencasting 6, as described above, is
truncated at its surface. The crushed rock material has been used for the building-up of the
nearby military fort.
Sodalite opencasting 7 (dyke B, S 17°20.890´, E 13°46.774´)
In the abandoned opencasting (120 m x 80 m) an up to 50 m thick carbonatite dyke is exposed
dipping with angles between 45° to 80° to the SW. The dyke is characterized by the dominance
51
of strongly tectonised anorthositic and syenitic xenoliths, surrounded by layered carbonatite
matrix (Fig. 3.1 d; 3.5 b; 3.6 b; 3.7 b; 3.10 d). Narrow fissures in the wallrock fragments are
invaded by carbonatitic material; sodalite is a minor constituent (up to 10 vol.%) or even absent.
The large diameter of the incorporated xenoliths (up to 2.3 m) and their high abundance in the
carbonatite dyke argue against a long distance of transport by the carbonatitic melt, but rather
suggest a passive emplacement of the carbonatite within previously tectonised shear-zones. A
conspicuous feature of opencasting 7 is the exposure of albite-biotite-magnetite-carbonate rock
remnants and schlieren within the carbonatitic breccia (Fig. 3.11 a, b), suggesting that its
intrusion was synchronous with or even predated the carbonatite emplacement.
Sodalite opencasting 8 (dyke A, S 17°20.470´, E 13°47.000´)
The open mine (50 m x 30 m) is the only opencasting that is presently mined. It is situated in the
central parts of a 26-30 m thick dyke of the sodalite-rich carbonatitic breccia (Fig. 3.9 a), dipping
approximately 45° to SW. Sodalite accumulations are mined bench-like as dimension stones of
45m3 in a first step, with the mining floor situated approximately 18 m below the surface level
(Fig. 3.9 b, c). For the export, these blocks are subsequently crushed to smaller blocks of 1 x 1 x
2 m.
In the central part of the carbonatite dyke sodalite is enriched in economic dimensions
(up to 85 vol.%), replacing feldspar-rich xenoliths of up to 1.5 m in diameter surrounded by or
intermingled with an ankerite-rich matrix (3.7 e; Fig. 3.9 d). Towards the dyke margins the
banded carbonatite variety dominates. The carbonatitic breccia is transsected by a younger
generation of almost silicate-free ferrocarbonatite veins and dykes (Fig. 3.12 a-d) and encloses
rare and irregularly distributed ultramafic xenoliths (UMX). Anorthosites in contact to the main
carbonatite dyke, exposed in the northern part of the opencasting, display a dyke-parallel
foliation with the elongated plagioclase crystals being partly replaced by sodalite (Fig. 3.2 d).
The changing modes of the anorthosite may be attributed to the interaction of hot carbonatitic
fluids with the bordering anorthosite.
In the eastern part of the opencasting a conspicuous WNW striking dark coloured and
sub-vertical dipping dyke of 1 m width is exposed that transsects the carbonatitic breccia (Fig.
3.11 c, d). The medium to fine-grained rock is characterized by a massive texture with albite,
biotite, ankerite and magnetite being the main constituents. It is invaded by coarse-grained
schlieren of the younger ferrocarbonatite generation. In the central part of the dyke welldeveloped magmatic flow textures can be observed. The dyke as well as the bordering lithologies
have been sampled in order to investigate their oxide and sulphide mineralogy. The results of the
52
investigation have been described by von Seckendorff & Drüppel (1999) and von Seckendorff et
al. (2000).
Fig. 3.9: Photographs of sodalite opencasting 8, which is the only opencasting that is presently mined (Locality: S
17°20.470´, E 13°47.000´). a) The mine is situated in the central parts of a large carbonatite dyke, crosscutting
leucogabbronorites of the dark anorthosite suite. b) and c) Sodalite-rich lithologies are mined bench-like, with the
sizes of the dimension stone blocks reaching up to 45 m3 in a first step. d) Sodalite is enriched in the central parts of
the carbonatite dyke. In these zones, sodalite (blue) replaces feldspar-rich xenoliths of up to 1.5 m in diameter, that
are in turn surrounded by and intermingled with ankerite-rich zones (yellowish).
3.3.6.1.3 Massive carbonatitic breccia [CB,m]
Twelve predominantly SE trending dykes of the massive variety of the carbonatitic breccia occur
in the area, with the single dykes being exposed over a narrow width of 0.5-15 m and a length of
70-450 m. In outcrop, the fine-grained to very fine-grained massive carbonatitic breccia is
characterized by a dark grey to black or brownish colour and a massive texture (Fig. 3.10 a, b).
Due to the small grain sizes individual minerals cannot be identified on a macroscopic scale,
however, the use of 5% HCl provides evidence for the presence of a certain amount of carbonate
minerals. In places, the massive carbonatitic breccia contains fragmented and pink-coloured
53
syenite xenoliths with diameters of 2 mm to 8 cm whereas anorthosite- and sodalite fragments
have never been observed on a macroscopic scale.
Narrow bands of the massive carbonatitic breccia may also occur interlayered with the
banded carbonatitic breccia (Fig 3.10 c). The massive layers display sharp boundaries parallel to
the dyke/wallrock interfaces, thus suggesting that the small grain sizes of the massive
carbonatitic breccia are generated by localised post-emplacement shearing and mylonitisation
rather than being a primary magmatic feature of this rock type. This interpretation is constrained
by the localised occurrence of massive and very fine-grained carbonatite lithologies in smallscaled shear zones (Fig. 3.10 d) of the carbonatitic breccia.
Fig. 3.10: a) On a macroscopic scale, the fine-grained massive carbonatitic breccia is generally characterized by a
dark grey to black colour and a homogeneous texture, with rare reddish syenite-fragments being observed locally
(Locality: S 17°20.066´, E 13°46.524´). b) Dykes of the massive carbonatitic breccia, displaying alternations of
massive brown, reddish and black coloured bands, most probably represent very fine-grained counterparts of the
layered carbonatitic breccia (Locality: S 17°20.147´, E 13°46.662´). c) Layers of the massive carbonatitic breccia
(dark grey) may also be intercalated with the layered carbonatitic breccia (light grey; Locality: S 17°20.147´, E
13°46.662´) or d) occur within small scaled shear zones within brecciated anorthosite (light grey) and syenite (pink;
Sodalite opencasting 7, locality: S 17°20.890´, E 13°46.774´).
54
3.3.6.2 K-feldspar-biotite-ilmenite-calcite rock
Rare irregular bodies of massive K-feldspar-biotite-ilmenite-calcite rocks (L) of up to 4 m in
diameter have been observed in several of the larger dykes of both the banded and the sodaliterich carbonatitic breccia (Fig. 3.11 a, b). Regarding their texture and macroscopic appearance,
these massive, dark coloured and medium-grained rocks strongly resemble both, the albitebiotite-rich fenite zones of the carbonatitic breccia and the so-called "black dyke" (Fig. 3.11 c,
d), which, however, never contain K-feldspar. Thus, the petrogenesis of the lamprophyre-like
rocks remained uncertain.
Fig. 3.11: a) and b) Remnants of the massive, black K-feldspar-biotite-ilmenite-calcite rock (black) are present
within a dyke of the sodalite-rich carbonatitic breccia (layered, with alternating grey albite-biotite-rich, blue
sodalite-rich and brown ankerite rich schlieren), thus suggesting that their emplacement predates the intrusion of the
carbonatites. (Sodalite opencasting 7, locality: S 17°20.880´, E 13°46.774´). c) and d) A grey dyke ("black dyke"),
mainly composed of albite, biotite, magnetite, and ankerite, transsects the sodalite-rich carbonatitic breccia (light
grey to bluish). This dyke is in turn invaded by younger ferrocarbonatite veins (brown; Sodalite opencasting 8,
locality: S 17°20.417´, E 13°47.000´).
55
3.3.6.3 Ferrocarbonatite
Both, the main carbonatite dykes and the lamprophyres are transsected by small veins and
stringers of a second generation of almost silicate-free ferrocarbonatite intrusions, up to 1.5 m in
width, which may be present in strongly varying proportions (Fig. 3.11 c, d; Fig. 3.12 a, b). The
medium to coarse-grained magnetite-ankerite rocks are characterized by a brownish surface
colour and a massive texture. Sub-rounded fragments of albite- and sodalite-rich xenoliths may
occur locally, concentrated in the central parts of the veins (Fig. 3.12 c), whereas trails of
euhedral magnetite grains mark the interfaces between the ferrocarbonatite veins and their
Fig. 3.12: Photographs of the late-stage ferrocarbonatite veins, exposed in the active opencasting 8 (Locality: S
17°20.470´, E 13°47.000´) a) Narrow, brown ferrocarbonatite veins and schlieren transsect and/or b) cement angular
fragments of the older carbonatitic breccia (light grey to bluish). c) The ferrocarbonatite veins (brown) suffered
minor wallrock contamination, when compared to the carbonatitic breccia (grey, pink, bluish). Xenoliths (white and
blue) are generally concentrated in the central parts of the veins. d) The ferrocarbonatite (brown) encloses
previously fenitized feldspar fragments (white to blue) but caused no metasomatic alteration of the wallrock by
itself. Trails of subhedral magnetite crystals (black) mark the contacts between the ferrocarbonatite veins and the
carbonatitic breccia.
56
wallrocks (Fig. 3.12 d). In contrast to the carbonatitic breccia, no effects of a metasomatic
alteration of the bordering rocks are visible on a macroscopic scale.
3.3.7 Dolerite [Dol]
Five predominantly SE striking dolerite dykes transsect the syenites, carbonatites and the
anorthositic rocks. The sub-volcanic rocks, exposed over a width ranging from 3 m to 70 m and a
length of up to 3.4 km, dominate in the eastern part of the study area. In natural outcrops the
rocks are characterized by thin brownish weathering crusts and spheroidal weathering. Outcrops
of solid bedrock are rare, mostly the fragments of bedrock are scattered as sub-rounded to
angular boulders with diameters continuously ranging from 20 cm to 2.5 m (Fig. 3.13 a).
Fig. 3.13: a) Dolerite dykes are commonly exposed as accumulations of scattered angular to sub-rounded, brownish
boulders (Locality: S 17°20.520´, E 13°47.380´). b) Fluvial terrace deposits border the Ondoto riverbed. The
carbonate-cemented immature conglomerate is mainly composed of well-rounded quartzite pebbles and gravel, with
sub-rounded to sub-angular quartz, anorthosite and carbonatite fragments being minor constituents of the reversely
graded terrace (Locality: S 17°21.060´, E 13°46.529´).
3.3.8 Trachyte [Tr] and Andesite [An]
In the northern part of the study area, a north-east trending carbonate-rich trachyte dyke is
exposed over 90 m in length and 3-15 m in width within the white anorthosite. The trachyte
differs from all other rock units investigated. It is characterized by spheroidal accumulation of
reddish K-feldspar crystals (~10 vol.%) with diameters of up to 4 cm and euhedral biotite
phenocrysts (~5 vol.%) set in a fine-grained dark-grey matrix mainly composed of carbonate, K-
57
feldspar, biotite, muscovite and euhedral magnetite. Due to the lack of exposed contacts between
the trachyte and the carbonatites, no temporal succession could be established.
Also exposed in the same area is an E-W striking and 3-8 m thick andesite dyke with a
length of 50 m. The grey andesite is characterized by narrow, reddish weathering surfaces.
Porphyritic plagioclase (An10-15), up to 2 mm in length, without a preferred orientation, is the
main mineral present (~60-70 vol.%) accompanied by minor biotite and amphibole as well as
accessory K-feldspar, magnetite and clinopyroxene. Since the andesite is solely exposed within
the white anorthosite its genetic relation to the other rock units of the study area remains
uncertain.
3.3.9 Sedimentary rocks
3.3.9.1 Unconsolidated deposits of fluvial terraces [Qtz,p]
Unconsolidated relics of fluvial terraces flank the ephemeral stream Ondoto. The variably eroded
quartzite pebble accumulations are exposed on different altitudes above the actual river floor and
thus allow the determination of their relative temporal succession.
The youngest generation of unconsolidated fluvial terrace deposits is exposed at an
altitude of 775-785 m above sea-level and is mainly composed of sub-rounded quartzite pebbles
(average: 80 %) with average diameters of 8 cm which mostly exhibit a reddish to brownish
weathering surface. In addition, variable amounts of subangular quartz fragments, up to 5 cm in
diameter, as well as fragments of epidote rocks (∅: 1-4 cm), anorthosite (∅: 0.2-0.5 m) and
carbonatite (∅: 0.1-0.4 m) may occur (Fig. 3.13 b).
Older fluvial terrace deposits occur at altitudes of 790-800 m and 805-818 m a.s.l.,
respectively. They differ from the younger terrace deposits by their higher amounts and larger
average diameters of sub-rounded to subangular quartzite pebbles (up to 95 %). In contrast,
small angular fragments of quartz, epidote rocks, anorthosite and carbonatite are rare. Single
quartzite pebbles, scattered on the steep ridges of the Zebra Mountains, have been observed in
altitudes of up to 915 m above sea-level.
In contrast to the epidote rocks, anorthosites and carbonatites, which are common rock
units of the study area, and the quartz-rich pegmatites which occur in close neighbourhood to the
sodalite mine, the quartzites are allochthonous rock material. The nearest occurrence of
58
quartzites of the Nosib Group, Damara supergroup, is actually situated approximately 40 km
south-west of the study area in the headwater area of the stream Ondoto. The changes in
composition and sizes of the three main fluvial terrace generations strongly suggest that the
sediments of the Nosib Group were continuously eroded during terrace formation with the
quartzite catchment area being displaced in further distance from the studied area. In addition,
sedimentary deposits of the Nosib Group occur in the north-eastern part of the Zebra Mountains
and thus strongly suggest that the whole anorthosite massif was once covered by the
Neoproterozoic sediments.
3.3.9.2 Carbonate-cemented deposits of fluvial terraces [Cgl]
The youngest fluvial terrace generation, exposed at the margins of the Ondoto riverbed at an
altitude of 743 to 752 m above sea level, is a carbonate-cemented immature conglomerate mainly
composed of well-rounded quartzite pebbles and gravel (∅: 1-15 cm). Sub-rounded to subangular quartz- (∅: 0.2-4 cm) and anorthosite (∅: 20-40 cm) fragments as well as carbonatite
anguclasts of up to 0.5 m in length are minor constituents (20-30%). The reversely graded terrace
has an average thickness of 2 m.
3.3.9.3 River floor deposits [All]
Both, the deposits of the ephemeral stream Ondoto and those of the numerous smaller blind
creeks are referred to as river floor deposits. The strath alluvial valleys are characterized by a
mostly plane river floor with strongly varying width depending on the stream-worn lithologies
and the orientation of the stream. Thus, the NE trending Ondoto river floor, mostly incising the
white anorthosite, is exposed over an average width of 300 m, whereas the width of the SE
trending creek beds, crossing the weathering-resistant dark anorthosites, ranges between 10-230
m.
The river floor of the river Ondoto is covered by quartz-rich and poorly sorted fine to coarse
sand together with minor sub-angular anorthosite, epidote and carbonatite fragments derived
from the surrounding rock units. In clear contrast, the river floor deposits of the blind creeks are
dominated by angular fragments of the closely neighbouring rocks.
59
3.3.9.4 Anorthositic talus deposits
Anorthositic talus deposits cover the steep mountains ridges in the north-western part of the
study area. The spheroidal anorthosite boulders with diameters of up to 10 m cover the
underlying dark anorthosite. The lack of hillwashes argues against a down-slope transport of the
anorthositic rock material but rather suggests an in situ formation of the talus deposits. The radial
drainage pattern of the Zebra Mountains causes the channelled eluviation of soil and thus causes
a belt-like distribution of the vegetation in regions of undisturbed soil formation.
3.3.9.5 Mining waste piles
Waste piles occur in the vicinity of the numerous abandoned and the active opencastings,
testifying to the continuous pressure-production history. In addition waste pile deposits have
been used as upbuilding for the settlement Orotumba.
60
3.4 Economic potential
3.4.1 Sodalite deposits
The sodalite deposits of Swartbooisdrif are mined since 1969 even though, recently, the turnover
of sodalite as a decorative stone is not without its problems. Because of the following reasons
only one of several opencastings (sodalite opencasting 8) is actually mined:
1)
Due to the quarry blasting during the early exploration activities, the sodalite-
bearing lithologies are heavily ruptured. Narrow fissures and cracks, mostly invisible on a
macroscopic scale, pass through all rock units exposed within the various opencastings. When
the sodalite-bearing lithologies are extracted, the rock pressure of the enclosing rock is lacking
and the cracks thus propagate fast until the whole block is penetrated and, as a result,
disaggregates. The speed of fissuration is even accelerated by day and night changes of
temperature, precipitous drops in temperature and chemical weathering.
2)
The external use of sodalite as well as its use as wall coverings of kitchens or
bathrooms is not practical since sodalite as a feldspathoid may be attacked by acid rain and weak
acids as commonly contained in cleaning agents. As a result, the sodalite exhibits a pale blue tint.
However, the carbonatitic breccia is not solely composed of sodalite but additionally contains
variable amounts of ankerite. Even though ankerite is characterized by a bright whitish colour on
fresh polished surfaces it is oxidised under atmospheric conditions after a certain time interval
and thus changes its surface colour from white into brown.
3)
The vast majority of superficial sodalite deposits have already been exploited
during the long-lasting pressure-production history. Thus the exploration activity has to be
expanded to deeper, not exposed levels of the carbonatite dyke. However, the sodalite
distribution within the dykes of the carbonatitic breccia is quite heterogeneous, as has been
mentioned before. In addition, the trends of the carbonatite dykes themselves are quite variable
including structural features as dying out, branching and curving. Following this the transfer of
the mining activity into greater depths bears risks.
As a conclusion the continuation of the exploration of the sodalite-rich lithologies as
dimension stones is not longer profitable regarding the high costs of mining, transport, export
and processing.
61
3.4.2 Leucotroctolites
In the Angolan part of the KIC medium-grained varieties of the black coloured leucotroctolites
of the dark anorthosite suite, commonly displaying iridescence of their plagioclase crystals, are
mined as ornamental stone, termed Angola Nero. In the Namibian part of the KIC, the
dominance of the white anorthosite suite as well as the rugged topography of the Zebra
Mountains argue against exploration activities concentrating on the dark anorthosite. However,
the leucotroctolites could be exploited by the sodalite mining company as a by-product without
causing considerable additional costs.
62
3.5 Tectonics
3.5 Tectonics
3.5.1 Tectonic overview
Following Trompette (1994), the study area is situated near the south-western margin of the
Proterozoic Congo Craton, which, particularly in its central parts, is covered by Cenozoic to
Quaternary sediments. Towards the north, south and west, the Congo Craton is bordered by four
major mobile belts, i.e. the Oubanguide fold belt in the north, the Damara fold belt in the south,
the Kaoko belt in the south-west and the West-Congo fold belt in the west, all of them being
formed during the Panafrican orogeny (Trompette, 1994). Towards the east, the Congo Craton is
limited by the Mesoproterozoic Kibara fold belt. Archaic to Palaeoproterozoic basement rocks
are mainly exposed in the marginal zones of the Congo Craton and are subdivided into four
major sub-zones (Trompette, 1994):
(1) the Cameroon-Gabon-Congo block in the north-west,
(2) the NE-Zaire block in the north-east,
(3) the Kasai block in the eastern centre and
(4) the Angolan block in the south-west.
Following this classification, the study area together with the Mesoproterozoic KIC and the
metamorphic basement rocks of the EC form part of the Proterozoic Angolan block. It can thus
be concluded that the EC, including the studied area, was stable part of the Angolan block during
Pan-Afican times and hence records the pre-Pan-African crustal evolution.
3.5.1.1 The anorthosite problem
As yet, no successful model has been established for the precise tectonic setting of the largescaled Proterozoic massif-type anorthosites, even though it is generally accepted that most of the
yet known anorthosite bodies are emplaced in extensional environments (Ashwal, 1993). By
now, a continental rift analogy is the preferred tectonic scenario which, however, has to differ
significantly from modern-day examples by both its topology and timing, since it has to account
for
(1) the ponding of large amounts of basaltic melts in the lower crust,
(2) an intrusion interval of several hundred million years,
(3) the linearity of most anorthosite massifs,
3.5 Tectonics
63
(4) the chemical bimodality and
(5) the apparent restriction of massif type anorthosite magmatism to the Proterozoic.
A plausible model has been presented by Hoffman (1989) who suggested that a stationary
supercontinent (Laurentia) was assembled during the early Proterozoic which acted as a thermal
blanket, leading to a thermal insulation of the mantle and thus producing a mantle "superswell".
It has to be tested, however, if a similar model may also account for the anorthosite magmatism
of the southern hemisphere.
3.5.1.2 The structural setting of carbonatites
Carbonatites, in general, are widely believed to be emplaced in the stable central parts of
continental areas. It has been demonstrated, however, that carbonatite magmatism may also be
linked to orogenic activity along plate margins. The following tectonic models were established
for carbonatite suites (see Woolley, 1989, for a review):
(1)
Rift valleys and lithosphere doming: Nearly half of the yet known occurrences are
emplaced within or nearby graben structures, as for example the Kaiserstuhl, Germany, and the
carbonatites of the East African rift. It has to be mentioned, however, that particularly the
carbonatite complexes of Southern Africa, including the Cretaceous carbonatites of Namibia and
Angola, are apparently linked to structural domes, with much of the faulting and rifting being a
response to the doming.
(2)
Major lineaments: Intrusions of carbonatites may be emplaced along major
lineaments ranging from ~10 km to several 1,000 km in length, as for example the Kapuskasing
group, Canada, the mid-Zambezi-Luangwa lineament, East Africa, and the intrusion lines of
Southern Africa, including Angola, Namibia and South Africa. Such lineaments, lying along and
above deep seated fracture zones, may persist over long periods of time and even predate the
carbonatite magmatism.
(3)
Orogenic activity: Carbonatite magmatism may also be correlated with orogenic
activity, with the carbonatites being emplaced in up to several 100 km wide zones adjacent to
orogenic belts, as for example the carbonatites of the North Atlantic area (Vertainen & Woolley,
1974).
Until now, the most successful tectonic model for the carbonatites of the Swartbooisdrif
area is their emplacement along major lineaments (model 2) which were already active before
their emplacement, as is supported by the coincidence of the orientation of faults intruded by
carbonatites with those used for the ascent of the older dark anorthosite melts. However, the
64
3.5 Tectonics
exact time span of anorthosite magmatism is yet unknown and may thus extend down to younger
ages. Following this, the faulting and subsequent carbonatite emplacement may also be a
response to structural doming induced by the intrusion of large masses of anorthositic melts, as is
inferred by model (1).
3.5.2 Structural investigations of the mapped sheet and related areas
If possible, the structural trends of all rock units exposed in the study area have been investigated
in some detail in order to constrain the temporal and spatial relationships between magmatism
and faulting. Structural investigations on anorthosite and granitoids of other parts of the
Namibian KIC as well as Landsat interpretations have been included in this study.
3.5.2.1 The Zebra Mountain massif
The KIC intruded as a generally N-S elongated intrusion in SW Angola, whereas it is exposed as
a more or less E-W trending body in Namibia. This sub-zone, which, in clear contrast to the
Angolan part of the KIC, is mainly composed of massive white anorthosite, most probably
represents an individual anorthosite intrusion. In the central part of this sub-zone, the white
anorthosite is transsected by predominantly ENE trending dark anorthosite sheets of up to 10 km
in width, which are responsible for the strong positive relief of the so-called Zebra Mountains
(Fig. 3.14 a, b). A similar orientation is displayed by a broadly N75E trending granite dyke
exposed near the southern margin of the KIC.
The Zebra Mountain massif extends over an area of approximately 1080 km2. Towards
the north-west, it is terminated by a N40E to N55E strike-slip fault bundle, up to 8 km in width,
along which the dark anorthosite is heavily tectonised (Fig. 3.15). This fault bundle, termed
Serpa-Pinto-Lineament, extends into Angola, where it is marked by intrusion-lines of granite,
separating the northern and southern part of the KIC, but is also apparent in the basement rocks
of the EC that border the KIC in the west. Thus, the Serpa-Pinto lineament can be interpreted as
an old structure, that already existed before the dark anorthosite and granite emplacement. In the
north-east, the Zebra Mountain massif is bordered by the S60E trending, fault-related river valley
3.5 Tectonics
65
Fig. 3.14: Photography of the Namibian KIC, as observed from a Cessna plane a) North of the Serpa-Pinto
lineament, the Namibian part of the KIC is dominated by the white anorthosite suite. Due to its low weathering
resistance, the deeply eroded white anorthosite generally forms low-lying, undulating planes. Rare plugs of dark
anorthosite are exposed as monadrocks (Locality: S 17°04.400´, E 13°13.380´, SE-direction). b) In clear contrast,
the steep ridges of the dark anorthosite suite dominate the Zebra Mountain massif south of the Serpa-Pinto
lineament. These rocks intruded the white anorthosite as generally E-W trending sheets (Locality: S 17°21.010´, E
13°21.129´, W-direction).
66
3.5 Tectonics
of the Kunene, whereas the tectonic contacts between the white anorthosite and the basement
mark the southern termination of the Zebra Mountains. The dark anorthosite sheets reach their
maximum width in the northern part of the Zebra Mountains as N65E to N75E trending ridges
and smaller, oval shaped plutons. The thickness of the sheets successively narrows from north
towards the southern margin of the KIC, where the dark anorthosite forms N70E to N85E
trending intrusion lines, up to 1 km in width and dipping 25-40° towards the SSE. A conspicuous
feature of the Zebra Mountains is the curving of the dark anorthosite sheets in the south-eastern
Fig. 3.15: Landsat image of the Namibian part of the KIC. Major lineaments and fault bundles are marked. Subset
of Landsat TM Scene 181-72 (RGB: 7-4-1).
3.5 Tectonics
67
and south-western parts of the massif, where the intrusion-lines strongly narrow towards their
tips and change their strike to run sub-parallel to the basement contacts.
The overall sheet-like shape of the Zebra Mountain massif makes a dyke-like ascent
mechanism appealing, however, several features argue against this process: (1) the sheet margins
are more curviplanar than is typical of dykes, (2) particularly in the central and marginal parts,
the sheets display geometries intermediate between those of dykes and diapirs, (3) no faults
extend in front of the narrow sheet tips. It is thus suggested, that the Namibian part of the KIC is
constructed of at least two diapiric pulses, with the dark anorthosite crystal mushes being
emplaced along pre-heated pathways. Following this, the sheet-like shape of the massif is mainly
controlled by the internal structure and general geometry of the Namibian part of the KIC. In
contrast, the overall shape of the E-W trending Namibian part of the KIC, including the white
anorthosite, may be influenced by pre-existing zones of structural weakness, inherited during a
Palaeoproterozoic collisional event, that affected the south-western margin of the Angolan block.
As yet, the most plausible three-dimensional expression of the Zebra Mountain massif is an ENE
elongated dome-like structure, as has already been proposed by Menge (1996). However, some
sort of NNW extension must have occurred in response to the doming as is evident from (1) the
general ENE trending orientation of both the dark anorthosite ridges and the elongated granite
intrusion at the southern margin of the KIC (Fig. 3.16 a and b) and (2) the dextral shear sense of
the NE trending Serpa-Pinto lineament, which was oriented oblique to the proposed 1,380 Ma
net opening (Fig. 3.16 c).
After its emplacement, the Zebra Mountain massif was subjected to extensive dextral
strike-slip faulting, that extends into the basement rocks of the EC in the south (Fig. 3.12). The
shear zones trend S70E in the western part of the Zebra mountains but continuously change their
strike towards S25E in the east. The strike-slip faults particularly displace the Serpa-Pinto
lineament, the white anorthosite and the granite intrusion. The horizontal displacement locally
reaches 3 km, giving the southern margin of the KIC a step-like appearance. In clear contrast, the
dark anorthosite sheets are not displaced abruptly along the shear zones but curve in the direction
of motion of the opposite block, giving the lineaments a drag fault appearance. The distortion of
the sheets strongly suggests that the dark anorthosite, in clear contrast to the white anorthosite,
still behaved in a ductile fashion during the displacement event, thus implying that the formation
of the dextral shear zones preceded the complete crystallization of the dark anorthosite. It can
thus be concluded that during late stages of cooling, the crystallising dark anorthosite sheets
within the KIC became capable of transmitting stress. It may be possible, that the faulting
reflects a regional stress field, characterized by an approximately E-W oriented maximum
68
3.5 Tectonics
horizontal extension, which became prominent as the repeated injection of anorthositic melt
came to a standstill. A regional stress-field like this would be in good accordance with the
general NNE trend of the Angolan part of the KIC, implying that opening was broadly E to SE,
oblique to the expected sheet-perpendicular extension. However, the bad exposure of the faults
complicated measurements of their dips and investigation of dislocation markers. Thus it was
impossible to reconstruct the precise stress-field, responsible for the faulting.
a Serpa-Pinto Lineament
b Dark Anorthosite Sheets
N
N
TIO
TA
LA
DI
N
ING
O PE N
25°
W
E
W
E
S
S
Rose Diagram Bi-Directional Serpa Pinto lineament
Total Number of Points = 15
Rose Diagram Bi-Directional Dark anorthosite
c Granite
d Joints in Dark Anorthosite
N
IN
OP EN
G
N
W
E
W
E
S
S
Rose Diagram Bi-Directional Granite
Rose Diagram Bi-Directional Dark anorthosite
Fig. 3.16: Rose diagrams, illustrating the trends of the ~1,380 Ma rock units and structures in the study area. Net
opening and dilatation are marked. a) Variation of strike of the dextral strike-slip fault bundle of the Serpa-Pinto
lineament. b) Variation of strike of the dark anorthosite sheets. c) Variation of strike of the fabric of the granite
intrusion. d) Variation of strike of joints in the dark anorthosite.
3.5 Tectonics
69
In contrast to the central part of the Zebra Mountains, the dark anorthosite plugs in the
Swartbooisdrif area, that represents a marginal zone of the Zebra Mountain massif, are mostly
irregularly bounded and rather display a irregularly scattered distribution within the white
anorthosite. The overall lack of extensive outcrops of solid bedrock prevents detailed structural
investigations of their fabric. However, the anorthositic rocks exposed in the Ondoto river valley
are transsected by numerous, SE trending fissures (Fig. 3.16 d). Joints in the rocks exhibit two
major orientations, i.e. (1) S10E, dipping 15-38° to the NE and (2) S50E, dipping 25-64° to the
SW. This structure does not copy a primary magmatic feature like extensive layering or igneous
lamination but rather represents secondary faulting, post-dating the emplacement of the
anorthositic magmas. Remarkably, the orientation of the joints correlates well with those of the
SE trending carbonatite dykes, suggesting that at least part of the carbonatites were emplaced
along pre-existing weakness zones within the anorthosite.
The structures in the anorthositic rocks of the Swartbooisdrif area are generally of brittle
origin, testifying to their late stage origin. Only anorthosites in direct contact to the dykes of the
carbonatitic breccia display a narrow zone of weak foliation marked by elongated plagioclase
and biotite crystals oriented sub-parallel to the dyke walls. This ductile deformation most
probably results from the local influx of heat and fluids released by the carbonatite melt.
3.5.2.2 The cogenetic syenite intrusions
Since the two syenite generations are genetically linked to, and were emplaced
contemporaneously with the dark anorthosite suite, a investigation of their structural orientation
should gain some general implications for the overall structural setting.
Measurements of the structural orientation of the first syenite generation are complicated,
since outcrops of solid bedrock are rare. Thus only 15 measurements are reliable, showing the
syenite to be emplaced as mainly E-W to S40E and minor S05E trending dykes, dipping 60-80°
to the NE, 65-89° to the SW and sub-vertically to E (Fig. 3.17 a, b). The varying trends of the
syenite dykes may reflect changes of the local stress-field, as discussed for the emplacement and
subsequent faulting of the dark anorthosite.
In contrast, the fenitized syenite, comprising the younger syenite generation,
predominantly trends S35E and S65E, mainly dipping between 45-60° to the SW and 60-75° to
the NE (Fig. 3.18 a, 3.19 a). These orientations correlate well with those of the late-stage strike-
70
3.5 Tectonics
slip fault bundle, exposed in the Zebra Mountain massif. Most likely, these weakness zones
provided excellent pathways for the emplacement of the younger syenite generation.
Both syenite dyke swarms are characterized by branching, dying out and curving of the
dykes as well as dyke segmenting into an en-échelon sets of intrusions towards their
terminations, suggesting minor changes in the maximum horizontal extension axis σ3 during
their emplacement. The brecciation of both syenite units is most probably related to repeated
movements along the dextral strike-slip faults.
a Syenite
b Syenite
N
N
W
E
S
Rose Diagram Bi-Directional Syenite
Lower Hemisphere Lambert (Equal Area)
Data Plotted as Poles Total Number of Points = 15
Fig. 3.17: a) Rose diagrams, illustrating the variation of strike of shearzones intruded by the ~1,380 Ma syenite. b)
Variation of dip direction and dip of the syenite dykes, marked as poles to planes.
3.5.2.3 The carbonatite dyke swarms
Dyke swarms typically form perpendicular to the least compressive stress σ3 and can thus be
used to constrain the patterns of, and the changes in the regional stress field. However, dykes
may also be injected along pre-existing fracture sets, opening in a direction oblique to their dyke
walls. The orientation of the carbonatite dykes in the area mapped reflects a major fracture set,
striking SE with a mean of 140°, thus implying a N50E trending, dyke-perpendicular maximum
horizontal extension axis (Fig. 3.18 a-d). However, a maximum dyke width is displayed by rare
S65E to S80E trending sodalite-rich carbonatites, which, at the same time, suffered negligible
post-emplacement shearing (Fig. 3.18 c). In clear contrast, the numerous carbonatite dykes
oriented oblique to this direction, i.e. S25E to S60E, contain weak ductile to brittle fabrics (3.18
3.5 Tectonics
71
a, b). Especially the massive carbonatitic breccia, trending S10E to S30E, was heavily sheared
parallel to the wallrock contacts after its solidification giving the very fine-grained rock a
mylonite-like appearance (3.18 d). Carbonatite dykes, oriented sub-perpendicular to the
dominant set and trending N to N20E, do also occur but are mostly undeformed.
a Layered Carbonatitic Breccia &
c Layered Carbonatitic Breccia
Fenitized Syenite
N
OP
EN
°
OP
E NIN
.5
G
27
L
DI
N
IO
AT
AT
W
20
ING
N
D
E
°
TIO
TA
I LA
N
W
E
S
S
Rose Diagram Bi-Directional
Fenitized syenite (and layered carbonatitic breccia)
Layered carbonatitic breccia
d Sodalite-rich Carbonatitic Breccia
b Massive Carbonatitic Breccia
N
EN
IN
OP
°
E
W
45
OP
EN I
N
G
G
N
N
TATIO
DIL A
W
S
S
Sdl-rich carbonatitic breccia
Massive carbonatitic breccia
E
Fig. 3.18: Rose diagrams, illustrating the trends of faults intruded by the ~1,380-1,340 Ma dykes of the fenitized
syenite (a) and the ~1,140-1,120 Ma carbonatite dykes (a-d). Net opening and dilatation are marked. a) Variation of
strike of faults intruded by both, the fenitized syenite and the carbonatitic breccia. b) Variation of strike of dykes of
the layered carbonatitic breccia. c) Variation of strike of dykes of the sodalite-rich carbonatitic breccia. d) Variation
of strike of dykes of the massive carbonatitic breccia.
These observations suggest that the dykes opened in a N20E direction, that was mostly
oblique to their walls. This is most likely to occur, if the carbonatite magma was injected in pre-
72
3.5 Tectonics
existing fracture sets, that were not perpendicular to the minimum horizontal extension axis at
the time of carbonatite emplacement. This interpretation is confirmed by several features of the
dykes, as for example (1) curving and branching of most of the dykes and (2) their dying out into
a set of en-echelon segments. Remarkably, the trends of the sheared carbonatite dykes are in
excellent agreement with those of the strike-slip faults, formed during or shortly after the
emplacement of the dark anorthosite suite, thus suggesting that most of the carbonatites were
injected along these older zones of weakness. It was impossible, however, to determine the
amount and direction of displacement of the sheared dykes, since the carbonatitic breccia is
generally characterized by a very inhomogeneous structure.
The dips of the carbonatite dykes are strongly variable, even within single dykes. In
general, however, the SE trending dykes of the carbonatitic breccia dip with average values of
45-60° to the SW and, subordinately, with average values of 60-75° to the NE, whereas the rare
NE trending dykes mainly dip between 55-65° to the NW and SE (Fig. 3.19 a-d). The general
coincidence of the dyke orientations of the ~1,380-1,340 Ma syenite and the ~1,140-1,120 Ma
carbonatite intrusions suggests that the two rock-units were emplaced along similar fracture sets,
which already formed during the final stages of anorthosite magmatism. The lineaments survived
a considerable time-span of >200 Ma, which makes an intervening orogenic event rather unlikely
and thus characterizes the time gap as an era of continental stability.
All dykes of the carbonatitic breccia display a variety of internal structures. The extensive
folding and layering of the carbonatitic breccia (Fig. 3.20a) have no preferred orientation and are
thus interpreted as magmatic features, as typically developed in high-viscosity melts. However,
the apparent high viscosity of the layered carbonatitic breccia is not a feature of the carbonatite
melt itself but is rather caused by the incorporation of wallrock fragments. These brecciated
and/or mylonitized xenoliths are present in strongly varying sizes and proportions. The large
amount and average size of the anorthositic and syenitic clasts make a transport by the
carbonatite unlikely but rather suggests that the carbonatite melts intruded previously sheared
faults.
Rafts of the layered carbonatitic breccia, oriented discordantly to the magmatic layering
of the surrounding carbonatite matrix (Fig. 3.20b), provide evidence for repeated injections of
carbonatite melt. During the subsequent injections of melt, the older, already crystallized parts of
the carbonatitic breccia suffered brittle deformation and displacement along small-scaled
fractures (Fig. 3.20 b, c).
3.5 Tectonics
73
a Layered Carbonatitic Breccia &
Fenitized Syenite
N
d Sodalite-rich Carbonatitic Breccia
N
c Layered Carbonatitic Breccia
N
b Massive Carbonatitic Breccia
N
Fig. 3.19: Variation of dip direction and dip of the dyke contacts of the ~1,380-1,340 Ma fenitized syenite (a) and
the ~1,140-1,120 Ma carbonatites (a-d), marked as poles to planes. a) Dip direction and dip of faults intruded by
both, the fenitized syenite and the carbonatitic breccia. b) Dip direction and dip of dykes of the layered carbonatitic
breccia. c) Dip direction and dip of dykes of the sodalite-rich carbonatitic breccia. d) Dip direction and dip of dykes
of the massive carbonatitic breccia.
Locally the carbonatites may also display features of ductile deformation, as for example
crystal-plastic deformation and re-crystallization of the matrix-carbonates. This holds true
especially for the massive carbonatitic breccia. In rare cases a weak magmatic lineation is
preserved in the carbonatitic breccia, marked by elongated mica, carbonate and magnetite
crystals (Fig. 3.20 d). This lineation has been investigated in some detail in the opencastings of
the two major dykes of the sodalite-rich carbonatitic breccia (dyke A and B; Fig. 2.1; Plate 2:
tectonic map), in order to constrain the flow pattern of the carbonatite melts. These
74
3.5 Tectonics
investigations demonstrate that the carbonatite melt generally followed a straight path of
upward-movement, rather than being injected obliquely to the fault planes.
Fig. 3.20: Photographs of common internal structures of the carbonatitic breccia. a) Impersistent magmatic folding
(Sodalite opencasting 4, locality: S 17°20.947´, E 13°47.752´). b) Rafts of the carbonatitic breccia (light grey to
bluish), wrapped by a younger generation of layered carbonatite (brownish; Sodalite opencasting 4, locality: S
17°20.947´, E 13°47.752´). c) Small-scale brittle deformation of already crystallized parts of the carbonatitic
breccia, surrounded by undeformed carbonate (brown) and silicate layers (grey, white, blue; Sodalite opencasting 4,
locality: S 17°20.967´, E 13°47.722´). d) Lineation (L) of the layered carbonatitic breccia, marked by elongate
ankerite crystals (Sodalite opencasting 3; locality: S 17°20.790´, E 13°47.685´).
3.5.2.4 Dolerites
The overall lack of well exposed possible dyke contacts allowed no detailed structural
investigation of this rock type. Measurements show that the dolerites were emplaced along the
same fracture set as the older intrusions of the carbonatitic breccia thus making major changes in
the regional stress field accompanying and/or predating their emplacement rather unlikely. The
3.5 Tectonics
75
dykes mainly strike SSE and dip with 45-75° to the WSW, but, in clear contrast to the massive
carbonatitic breccia, show no effects of syn- to post-emplacement shearing.
3.5.3 Fault system
Faults dominate in the southern part of the area mapped. They are exposed as mainly NE and
subordinately SE trending fault bundles, mostly intruded by syenite. The fractures are mostly
filled by massive epidote rocks, replacing the former syenite and/or cementing angular syenite
clasts. Fault planes are never exposed but the conspicuous green colour of the epidote
mineralization allowed to follow the path of the faults over a certain distance. However, the
direction and value of fault-related movement could only be estimated by the relative
displacement of the syenite, carbonatite and dolerite dykes. A method like this, however, bears
risks, since it remains uncertain if the dyke displacement reflects a post-intrusive tectonic event.
This holds true especially for most of the carbonatite and dolerite intrusions, which have been
shown to be emplaced along a complex pre-existing fracture system.
Indeed, the various dykes appear to be displaced sinistrally and dextrally across both the
NE and the SE trending faults, suggesting that this separation is the result of a jump-over to other
faults. This interpretation is constrained by a dolerite dyke, that apparently suffered a sinistral
displacement of 200 m, which, however, has not affected the neighbouring older syenite and
carbonatite dykes.
76
4 General remarks
77
PART II: Petrogenesis and evolution of the Kunene Intrusive
Complex and the spatially associated felsic suite
4 General remarks
Massif-type anorthosite complexes, commonly associated with minor felsic rocks, are
characteristic features of the Proterozoic crust. Their apparent temporal restriction to the
Proterozoic testifies to specific tectono-thermal regimes and crust-forming processes that were
operative during this period. However, the petrogenesis of Proterozoic anorthosite complexes is
still a matter of debate. Thus, no general agreement has been reached on issues such as formation
of parental magmas, monobaric vs. polybaric crystallization, mode and depth of emplacement
and cooling history (e.g. Ashwal, 1993, for a review). Anorthosites, leucotroctolites and related
gabbroic rocks commonly display late-magmatic and subsolidus reaction textures, among which
garnet-bearing coronas are most conspicuous (e.g. Griffin & Heier, 1973; Whitney &
McLelland, 1973; McLelland & Whitney, 1977, 1980a, b; Johnson & Essene, 1982; Bingen et
al., 1984; Griffin et al., 1985; Grant, 1988; Johnson & Carlson, 1990; Ashwal et al., 1998). These
textures may place constraints on at least two of the manifold problems pertaining to the
petrogenesis of massif-type anorthosite complexes, namely (i) their intrusion depth, i.e. lower vs.
upper crustal level, and (ii) their subsolidus evolution, i.e. slow isobaric cooling vs. a separate
prograde metamorphic overprint.
As summarized by Ashwal (1993), the intrusion of massif-type anorthosites into the
lowermost crust (e.g. Martignole & Schrijver, 1971; Taylor et al., 1984) is no longer generally
accepted. Based on various evidence, a shallow intrusion depth has been invoked, for example
for the Nain anorthosites of Labrador (Berg, 1977a, b), the Laramie anorthosite of Wyoming
(Fuhrmann et al., 1988; Kolker & Lindsley, 1989; Grant & Frost, 1990) or for the Adirondack
anorthosites of New York (Valley & O’Neil, 1982). A process of slow isobaric cooling for the
subsolidus formation of corona structures was already proposed by Martignole & Schrijver
(1971), Griffin (1971) and Bingen et al. (1984). In other cases, however, an alternative model,
considering a separate event of prograde granulite facies metamorphism seems to be justified
(McLelland & Whitney, 1977, 1980a, b; Johnson & Essene, 1982).
Most anorthosite massifs contain a small volume of temporally and spatially associated
felsic plutonic rocks, characterized by a composition ranging from diorite to granite and a mostly
anhydrous mineralogy. Commonly, intrusive contact relationships suggest these granitoids to be
78
4 General remarks
younger than the associated anorthosite (e.g. de Waard, 1970; Seifert, 1978). The fact, that the
mineral chemistry of the Fe-Mg silicates and plagioclase in the silicic suites appears to extend
the compositional ranges of the anorthositic minerals to higher XFe and lower An-contents, has
been interpreted in terms of consanguinity between the two rock suites (e.g. Fuhrman et al.,
1988). An alternative model considers the granitoids to represent a chemically independent,
crustal derived intrusive suite (e.g. Anderson, 1980; Anderson & Bender, 1989; Emslie, 1991;
Emslie et al., 1994).
Focusing on these pairs of alternative models, the geochemical signatures as well as the
magmatic and subsolidus reaction textures recorded by the anorthositic rocks of the Kunene
Intrusive Complex and the spatially associated granitoid suite, have been investigated in some
detail. Parts of this study have already been presented in Drüppel (1999) and Drüppel et al.
(2001).
5 Petrography
79
5 Petrography
5.1 Petrography of the anorthositic rock suite
5.1.1 Igneous textures
All anorthositic rock types are characterized by a massive appearance with coarse-grained
plagioclase crystals without preferred orientation. The grain sizes of plagioclase commonly vary
between 0.2-5 cm. Both, the white and the dark anorthosite suites display the typical magmatic
assemblage of
plagioclase ± orthopyroxene ± clinopyroxene + ilmenite + magnetite + biotite + amphibole ±
apatite ± zircon.
In contrast to the white anorthosite suite, the dark anorthosites may additionally contain
variable amounts of olivine.
5.1.1.1 White anorthosite suite
Eight samples of the white anorthosite from the sodalite mining area have been investigated in
detail.
White anorthosite (A,w):
Ku-97-08b, Ku-97-33, Ku-97-33a, Ku-97-33b, Ku-97-44, Ku-97-99a,
Ku-98-45, Ku-98-60
A characteristic feature of the white anorthosite is the dominance of anhedral cumulus
plagioclase (87-99 vol.%), which, however, suffered an almost complete hydrothermal alteration
(i.e. saussuritization, albitization and/or sericitization), responsible for the white to greenish
colour of these rocks. Interstitial ortho- and clinopyroxene are minor phases. Late-magmatic
biotite and pargasite form discontinuous rims around clinopyroxene and anhedral ilmenite.
Olivine or its alteration products have never been observed in the investigated white anorthosite
samples.
Due to a strong brittle deformation, plagioclase, pyroxene, amphibole and biotite are in
most cases heavily brecciated. Locally, plagioclase is recrystallised to granular mosaics. Cracks
in the rocks are filled by carbonate and/or epidote.
80
5 Petrography
Anhedral, tabular plagioclase is virtually completely sericitized (Fig. 5.1a) or altered to
fine-grained intergrowths of albite, epidote, actinolite and calcite (saussuritization, Fig. 5.1b).
Plagioclase grain size is mostly coarse-grained with an average of 2.3 cm. Non-altered, twinned
(albite-, pericline-, Carlsbad-law) plagioclase relics are preserved in the central parts of large
plagioclase grains (Fig. 5.1a). The normally to slightly oscillatory zoned magmatic plagioclases
fall in the compositional range of An43-53 (andesine to labradorite). However, with an average of
2 mol.%, the differences in rim and core compositions are negligible. Locally, the primarymagmatic plagioclase crystals are recrystallised to granular mosaics (Fig. 5.1c). A characteristic
feature of these secondary plagioclase crystals is the general lack of alteration and twinning
(except rare albite-twins) and an inverse zonation, with An3-4 cores and An5-10 rims.
Intercumulus orthopyroxene (XMg: 0.71) is partially to completely replaced by chlorite.
The fine-grained (grain size: < 0.5 mm), subhedral grains commonly display ilmenite exsolution
lamellae  (100) and are surrounded by narrow and impersistent chlorite and/or pargasite rims.
Anhedral to subhedral intercumulus clinopyroxene is generally fine-grained (grain size: < 1mm)
and heavily tectonised. Due to a hydrothermal overprint, the granular clinopyroxene grains are
completely replaced by oriented and fine-grained actinolite intergrowths (Fig. 5.1 d).
Clinopyroxene is enveloped by a narrow rim of late-magmatic, green pargasite.
Ore minerals, i.e. magnetite and ilmenite, are minor constituents of the white anorthosite.
They occur as discrete anhedral grains in the interstices of cumulus plagioclase, suggesting their
late-magmatic origin. However, subhedral magnetite may additionally be included in
plagioclase. Commonly, ilmenite displays polysynthetic twinning. Both ore-minerals underwent
subsolidus deformation along small-scaled shear zones (Fig. 5.1 e).
Light green to colourless actinolite replaces clinopyroxene, whereas late-magmatic green
pargasite (XMg: 0.65-0.70), together with biotite (XMg: 0.48), forms narrow rims around
intercumulus orthopyroxene, clinopyroxene and ilmenite (Fig. 5.1 e). Marginally, the biotite
grains are replaced by fine-grained intergrowths of chlorite, calcite and epidote. Subhedral,
prismatic apatite inclusions (average grain size: 0.1 mm) in clinopyroxene and ilmenite are a
common accessory phase.
Chlorite and epidote replace biotite and plagioclase and additionally fill narrow fissures
in the rocks. Intergranular calcite may be present in significant concentrations, reaching up to 4
vol.%.
5 Petrography
81
Fig. 5.1: Polished-section photomicrographs of the white anorthosite suite. Cross-polarized light (a-d). Planepolarized light (e-f). a) In rare cases the cores of lathshaped cumulus plagioclase are preserved whereas the rims are
pervasively sericitized. b) Saussuritisation of cumulus plagioclase. c) Granoblastic recrystallisation of cumulus
plagioclase to almost pure albite. d) Pseudomorphous replacement of clinopyroxene by actinolite. e) Deformation
along micro-shearzones causes the recrystallisation of late-magmatic biotite, surrounding interstitial ilmenite. f) A
xenocryst of relict andradite-rich garnet, that is marginally replaced by retrograde chlorite, is surrounded by rims of
late-magmatic amphibole.
Relic xenocrysts of andradite-rich garnet (average composition: And72Grs27), present in
two of the nine investigated samples of white anorthosite, point to a contamination of the
parental melt with calc-silicate country rock. The tectonised garnet grains are partially replaced
by chlorite and surrounded by late-magmatic pargasite (Fig. 5.1f).
82
5 Petrography
5.1.1.2 Dark anorthosite suite
In the investigated 59 dark anorthosite samples coarse-grained plagioclase is the dominant
mineral (65-99 vol. %) with mostly anhedral to subhedral, tabular shape. Dust-like inclusions of
Fe-Ti oxides in plagioclase are common in all samples of dark anorthosites but are most
abundant in the leucotroctolites and thus most probably account for their dark-coloured
appearance. A common feature is mortar texture, consisting of primary plagioclase, surrounded
by re-crystallized grains of similar composition, possibly caused by the movement of the
anorthositic crystal mushes (e.g. Longhi & Ashwal, 1985; Morse, 1968).
Minor post-cumulus orthopyroxene (En59-63) and clinopyroxene (diopside with XMg: 0.660.76) are the dominant minerals (3-12 vol.%) among mafic silicates in the leucogabbro and norite samples. They occur as relatively large, subhedral to anhedral intercumulus grains in
subophitic relationship with cumulus plagioclase. In the leucotroctolite subordinate olivine (5-8
vol.%; Fo54-65) occurs as small subhedral inclusions in plagioclase or as anhedral grains in the
interstices between plagioclase crystals, partially rimmed by ortho- or clinopyroxene (up to 3
vol.%). The heavily strained intercumulus olivine is partially to completely replaced by finegrained intergrowths of talc, carbonate and quartz, together with magnetite, forming irregular
vein fillings in cracks of olivine. In contrast to the leucogabbro- and –norite and the
leucotroctolite, anorthosite sensu stricto contains only minor amounts of pyroxene and/or
olivine (0.5-3 vol.%).
Common accessories of all lithologies of the dark anorthosite suite are anhedral ilmenite
and magnetite, with curved grain boundaries. The Fe-Ti ores form discrete grains in the
interstices between plagioclase crystals, indicating a post-plagioclase formation. However,
magnetite may additionally occur as fine-grained inclusions in plagioclase and olivine, pointing
to its early crystallization. Intercumulus ilmenite, ortho- and clinopyroxene are partially
surrounded by narrow rims of intergrowths of biotite (XMg: 0.39-0.67) and brown to yellowish
amphibole (XMg: 0.55-0.67). In the leucogabbros and -norites amphibole and biotite of similar
compositions also occur localised in patches, bearing abundant inclusions of fine-grained,
euhedral apatite.
The well preserved magmatic textures as well as the analysed compositions of the
magmatic mineral assemblages in the investigated leucotroctolite and anorthosite samples of the
dark anorthosite suite resemble those reported for the Angolan part of the KIC (Ashwal & Twist,
5 Petrography
83
1994; Carvalho & Alves, 1990; Morais et al., 1998; Silva, 1990, 1992; Simpson, 1970; Vermaak,
1981).
5.1.1.2.1 Pyroxene-bearing anorthosite and leucogabbronorite
Nine samples of pyroxene-bearing anorthosites and 25 samples of leucogabbronorites
(comprising 1 sample of leucogabbro and 24 samples of leucogabbronorite s.s.) were taken from
the Swartbooisdrif mine and the area around the settlement Swartbooisdrif.
Pyroxene-bearing anorthosite (A,px): Ku-97-02, Ku-97-12, Ku-97-13, Ku-97-31, Ku-97-92a, Ku97-93, Ku-98-78, Ku-98-95, Ku-98-123
Leucogabbronorite s.s. (GN): Ku-97-03, Ku-97-03b, Ku-97-03c, Ku-97-04, Ku-97-05, Ku-97-06,
Ku-97-08a, Ku-97-30, Ku-97-31b, Ku-98-71, Ku-98-72, Ku-98-79, Ku-98-84, Ku-98-92,
Ku-98-126, Ku-98-128, Ku-98-129, Ku-98-222, Ku-01-19, Ku-01-20, Ku-01-21, Ku-01-22,
Ku-01-23, B-98-101a
Leucogabbro (GN): Ku-98-68
Tabular plagioclase, with an average grain size of 1.8 cm, is the major constituent (65-98 vol.%.)
of the rock samples investigated (Fig. 5.2 a-c). Subordinate primocrysts of lath-shaped
plagioclase have been observed in four samples (Fig. 5.2 f). The majority of the analysed
plagioclase crystals (An43-54) shows minor chemical oscillatory or patchy zoning. Marginally, the
plagioclase crystals are recrystallised to medium-grained crystals (average grain size: 5 mm) of a
similar compositional range. In one sample (Ku-98-84) plagioclase crystals adjacent to
orthopyroxene display thin strongly reversed rims where the An content increases abruptly from
54 mol.% to 69 mol.%. According to Morse & Nolan (1984), these more calcic rims cannot be
explained by subsolidus processes, but require an origin from intercumulus liquid. In contrast to
the white anorthosite, plagioclase of the leucogabbronorites and the pyroxene-bearing
anorthosites of the dark anorthosite suite suffered only minor sericitization, which occurs
localised along small cracks or poicilitic magnetite inclusions. Alteration products formed by
saussuritization are rare.
Minor post-cumulus orthopyroxene and clinopyroxene (XMg: 0.53-0.65), both lacking
significant zoning, are the dominant minerals among mafic silicates. They occur as relatively
84
5 Petrography
large (up to 1.5 cm), anhedral intercumulus grains (Fig. 5.2 a-c) or in subophitic relationship
with cumulus plagioclase (Fig. 5.2 f). Clinopyroxene is commonly partially to completely
replaced by colourless to light green, fibrous actinolite and surrounded by rims of late-magmatic
Fig. 5.2: a) Polished-section photomicrographs of the pyroxene-bearing anorthosites and the leucogabbronorite.
Cross-polarized light (a-c, f). Plane-polarized light (d-e). a) Uralitised interstitial clinopyroxene fills the interstices
between heavily sericitized cumulus plagioclase, and is surrounded by a rim of late-magmatic amphibole. b) Former
pyroxene is replaced by epidote, whereby ilmenite-exsolution lamellae are preserved. c) Intercumulus
orthopyroxene is commonly completely altered to chlorite and rimmed by late-magmatic amphibole. d) Brownish,
strongly pleochroitic biotite and amphibole aggregates surround anhedral, interstitial ilmenite grains. e)
Poikiloblastic amphibole bears abundant apatite inclusions, both of them crystallized from trapped interstitial
liquids. f) Subophitic texture displayed by interstitial, uralitized clinopyroxene and lath-shaped, subhedral
plagioclase.
5 Petrography
85
amphibole (Fig. 5.2 a). In clear contrast, orthopyroxene grains are always completely replaced by
chlorite and/or epidote, whereby previously exsolved platelets of ilmenite, oriented || (010) and ||
(001), are preserved (Fig. 5.2 b-c).
Ubiquitous intercumulus ortho- and clinopyroxene as well as ilmenite, are partially
surrounded by narrow rims of intergrowths of biotite (XMg: 0.39-0.48; Fig. 5.2 d) and greenish to
brownish amphibole of mostly (ferro-)pargasitic to (magnesio-)hastingsitic composition (XMg:
0.46-0.73; Fig. 5.2 a). Amphibole and biotite of similar composition also occur localised in
patches, bearing inclusions of fine-grained, euhedral apatite (Fig. 5.2 e). Exsolution of Ti-phases
like ilmenite or rutile has not been observed in the amphiboles, indicating that they did not reequilibrate with respect to Ti. The observed textural positions and compositions of biotite and
amphibole are consistent with growth from pockets of interstitial magma trapped between the
cumulus minerals and indicate that pargasite and biotite crystallized late in the magmatic history.
In contrast, subordinate green to colourless, fibrous actinolite and magnesio-hornblende (Al:
0.27-0.98; Ti: < 0.04; XMg: 0.60-0.65) was formed by the alteration of intercumulus ortho- and
clinopyroxene.
Anhedral intercumulus ilmenite with embayed grain boundaries (Fig. 5.2 d) is the
dominant mineral among Fe-Ti oxides, whereas anhedral magnetite with ilmenite exsolution
lamellae, oriented ||{111} and {100}, is rare. Common accessories are zircon, late-magmatic
apatite as well as the alteration products epidote (Ps: 0.14-0.28), chlorite (XMg: 0.43-0.62),
clinozoisite, calcite and sericite.
In the leucogabbronorite sample Ku-97-03, a xenocryst of a light pink andraditic garnet
has been observed, which resembles the garnet contained in the white anorthosite sample Ku-9708b, regarding both its composition and textural position. It can thus be concluded that the
leucogabbronorite/anorthosite units of the dark anorthosite suite, like the older white anorthosite
suite, underwent contamination by crustal calc-silicate lithologies during their uprise.
5.1.1.2.2 Olivine-bearing anorthosite and leucotroctolite
Eight samples of olivine-bearing anorthosite as well as 17 samples of leucotroctolite have been
taken from the sodalite mining area and from the southern, northern and western margins of the
Zebra mountains.
86
5 Petrography
Olivine-bearing anorthosite (A,ol): Ku-97-28, Ku-97-28a, Ku-97-95, Ku-98-23, Ku-98-124, B-98274a, B-98-274b, B-98-281b
Leucotroctolite (T): Ku-97-92, Ku-97-96, Ku-97-104, Ku-97-105, Ku-98-41, Ku-98-52, Ku-98-53,
Ku-98-77, Ku-98-94, Ku-98-121, Ku-98-125, Ku-98-220, Ku-98-221a, Ku-98-221b, Ku98-228, Ku-98-230, Ku-98-231
Subhedral to anhedral, tabular plagioclase with embayed grain boundaries is the main mineral
present (75-96 vol.%). The crystals have an average grain size of 5 mm, which, however,
strongly decreases in the vicinity of intercumulus pyroxene and olivine (Fig. 5.3 a). The majority
of the weakly zoned plagioclase falls in the compositional range An47 to An67, whereas reversed
rims of bytownitic composition may occur locally. Common features of plagioclase of the dark
coloured leucotroctolite suite are (1) the overall presence of dust-like particles of Fe-Ti oxides
(Fig. 5.3 b), (2) a weak sericitization and (3) the general lack of saussuritization products (Fig.
5.3 a-f).
Subordinate olivine (Fo54-65) occurs as small subhedral inclusions of up to 0.2 mm in
diameter in plagioclase or as anhedral grains (average grain size: 4 mm; Fig. 5.3 c) in the
interstices between plagioclase crystals, partially rimmed by orthopyroxene, clinopyroxene
and/or biotite (Fig. 5.3 d). Both olivine types may contain rounded magnetite inclusions (Fig. 5.3
c). In most cases, the heavily strained intercumulus olivine is partially to completely replaced by
fine-grained intergrowths of talc, carbonate and quartz, together with magnetite, forming
irregular vein fillings in cracks of olivine. As no relic serpentine mineral was observed, the
replacement most probably is a direct reaction of the fayalite component to magnetite, producing
the SiO2 needed to transform the forsterite component to talc:
3 Fe2[SiO4] + O2 → 2 Fe2+Fe3+2O4 + 3 SiO2
Fa
Mag
(5-1)
Qtz
2 Mg2[SiO4] + 2 SiO2 + H2O + CO2 → Mg3[Si4O10/(OH)2]+ MgCO3
Fo
Qtz
Tlc
(5-2)
Mgs
This replacement is the result of a weak hydrothermal alteration of the dark anorthosite suite.
Anhedral clinopyroxene (XMg: 0.60-0.74), orthopyroxene (XMg: 0.65-0.70), ilmenite and
magnetite, with curved grain boundaries, form discrete grains in the interstices between
5 Petrography
87
plagioclase crystals, indicating a post-plagioclase formation (Fig. 5.3 e). Clinopyroxene and
orthopyroxene, together with biotite, may also form thin irregular rims around interstitial olivine
(Fig. 5.3 a, d). Magnetite may additionally occur as fine-grained, rounded inclusions in
plagioclase and olivine (Fig. 5.3 b-c), pointing to an early crystallization of magnetite.
Fig. 5.3: Polished-section photomicrographs showing common magmatic textures of the olivine-bearing
anorthosites and leucotroctolites. Cross-polarized light (a-c, e-f). Plane-polarized light (d). a) Interstices between
coarse-grained cumulus plagioclase are filled by interstitial olivine, which in turn is surrounded by clinopyroxene
rims. b) Fine-grained inclusions of cumulus magnetite in plagioclase account for the dark macroscopic colour of the
leucotroctolites. c) Olivine may occur as sub-rounded inclusions in cumulus plagioclase, testifying to its early
formation. d) Subhedral, interstitial olivine is surrounded by irregular rims of orthopyroxene and biotite, filling the
open spaces between plagioclase crystals. e) Anhedral clinopyroxene with ilmenite-exsolution lamellae may form
discrete intercumulus grains. f) Plagioclase encloses annealed cracks, filled by amphibole-clinopyroxene
symplectites.
88
5 Petrography
Ilmenite, spinel, clinopyroxene and olivine are partially surrounded by narrow rims of
late-magmatic biotite (XMg: 0.45-0.48) and rare magnesio-hastingsitic to pargasitic amphibole
(XMg: 0.77-0.86). Common accessories are epidote and chlorite (XMg: 0.71)
Small cracks in the rocks are filled by amphibole-clinopyroxene symplectites, that are locally
overgrown by re-crystallized plagioclase (Fig. 5.3 f).
5.1.2 Structures of subsolidus origin
In five of the 17 leucotroctolite samples of the dark anorthosite suite (Ku-97-104, Ku-97-105,
Ku-98-220, Ku-98-221a, Ku-98-221b), the igneous mineral assemblages and textures are locally
obliterated by subsolidus reactions, leading to the development of two different types of garnetbearing corona textures, surrounding olivine and ilmenite, respectively.
In addition, conspicuous ilmenite-magnetite-orthopyroxene symplectites developed in the
vicinity of olivine in five anorthosite and leucotroctolite samples (Ku-97-52, Ku-97-104, Ku-9857, Ku-98-79, Ku-98-221b). Both types of corona textures have been investigated in some detail
by Drüppel et al. (2001).
5.1.2.1 Type I corona
In five leucotroctolite samples, olivine in contact with plagioclase (An59-75) is rimmed by blocklike, euhedral orthopyroxene (En62-71) which, in turn, is surrounded by isolated euhedral garnet
grains (average composition: Prp28.6Sps3.4Alm54.9Grs8.6And4.5; XMg: 0.31-0.42) (Fig. 5.4 a-c).
Subordinate quartz may appear in small grains next to garnet. In a simplified way, the observed
garnet-forming reaction can be written as
Ca[Al2Si2O8] + (Fe, Mg)2 [Si2O6] → Ca (Fe, Mg)2 [Al2Si3O12] + SiO2
An
Opx
Grt
(5-3)
Qtz
Using the actual mineral compositions, the reaction cannot be balanced isochemically,
but would lead to excess Na, Ca and Si in the reaction products. At the contacts with plagioclase,
garnet and orthopyroxene are replaced by fibrous green to colourless pargasite to magnesiohastingsite (Al: 2.73-3.15; Ti: < 0.01; XMg: 0.77-0.87) (Fig. 5.4 a-c) or biotite (XMg: 0.61-0.75).
5 Petrography
89
Block-like orthopyroxene grains surrounding olivine may contain small inclusions of vermicular
magnetite (Fig. 5.4 d) but no platelets of ilmenite. This feature can be attributed to an increase in
oxygen fugacity and/or cooling in the presence of a fluid phase, following the balanced reaction
1.13 Mg1.22Fe0.78[SiO4] + 0.35 O2 + 0.87 SiO2 → Mg1.38Fe0.62[Si2O6] + 0.083 Fe2+Fe3+2O4
(5-
4)
Ol
Opx
Mag
In contrast to the irregular, anhedral orthopyroxene rims around olivine and interstitial
orthopyroxene grains, which are presumably of igneous origin, the block-like orthopyroxene
crystals of the corona structures are most probably due to subsolidus reactions between
plagioclase and olivine or to recrystallisation of magmatic orthopyroxene rims. It should be
noted, however, that there is no significant compositional difference between igneous
orthopyroxene grains and those of subsolidus origin. This fact is taken as evidence that, during
the subsolidus stage, both types of orthopyroxene were thoroughly re-equilibrated.
5.1.2.2 Type II corona
Two of the six leucotroctolite samples (Ku-98-57, Ku-98-221b) additionally contain reaction
rims of euhedral garnet (average composition: Prp22.7Sps3.3Alm59.2Grs10.6And4.2; XMg: 0.23-0.30),
associated with subhedral, tabular biotite (XMg: 0.66-0.77), around magmatic ilmenite or
magnetite with lamellar ilmenite oxy-exsolution. Both garnet and biotite are in direct contact
with the Fe-Ti oxide (Fig. 5.4 e-f), whereas amphibole has never been identified in these corona
structures, even though it commonly occurs as late-magmatic rims around ilmenite in all other
samples. These textural features suggest a garnet-forming reaction involving the anorthite
component of igneous plagioclase, the Tschermak’s component of the late-magmatic pargasite
and the Fe component of an igneous Fe-Ti oxide, possibly a titano-magnetite:
Ca[Al2Si2O8] + Ca2(Fe, Mg)3Al2[Al2Si6O22] (OH)2 + Fe3O4 + SiO2 →
An
Ts in Am
3 Ca(Fe, Mg)2[Al2Si3O12] + H2O + ½O2
Grt
Mag
Qtz
(5-5)
90
5 Petrography
In reality, this reaction is much more complex and presumably allochemical. Compared to the
type-I corona textures, such a reaction mechanism is supported by the slightly lower XMg of
garnets in type-II coronas, which may be the consequence of Fe-Ti oxide participation in the
garnet forming reaction (5-5).
Fig. 5.4: Polished-section photomicrographs of garnet-bearing corona textures, developed in the leucotroctolites of
the Kunene Intrusive Complex. Plane-polarized light. a) Olivine, partially replaced by subsolidus orthopyroxene,
which in turn is surrounded by garnet. In contact to plagioclase, both garnet and orthopyroxene are altered to fibrous
actinolite. b) Rim of block-like orthopyroxene, surrounded by euhedral garnet grains and fibrous actinolite. c)
Partially corroded garnet grains, surrounded by actinolite as an alteration product. d) Block-like orthopyroxene,
intergrown with fine-grained magnetite surrounds altered olivine. e) Euhedral garnet, bearing abundant biotite
inclusions and being intimately associated with tabular biotite surrounds igneous ilmenite. f) Garnet-biotite corona
around orthomagmatic ilmenite. Both, garnet and biotite are in direct contact to ilmenite.
5 Petrography
91
5.1.2.3 Ilmenite-magnetite-orthopyroxene symplectites
In five more leucotroctolite samples, one of them containing type-I corona structures,
conspicuous ilmenite-magnetite-orthopyroxene symplectites (IMOS) developed between
intercumulus olivine (Fo54-65) and magmatic Fe-Ti oxides.
Both the symplectites and the bordering olivine are enclosed by a continuous rim of
block-like orthopyroxene (Fig. 5.5 a) that also separates the symplectites from the magmatic FeTi oxide grains. The IMOS probably formed at the expense of olivine and magmatic Fe-Ti
oxides, following the balanced reaction
1.05 Mg1.22Fe0.78[SiO4] + 0.015 TiO2 +0.85 SiO2 + 0.133 O2 →
Ol
Mg1.28Fe0.71[Si2O6] + 0.028 Fe2+Fe3+2O4 + 0.015 FeTiO3
Opx
Mag
(5-6)
Ilm
Since the reaction was not isochemical, the formation of ilmenite in the symplectites is
only possible if Ti is introduced into the system. This may explain the presence of symplectites
in close proximity to magmatic ilmenite. Where olivine is in contact with the symplectite bodies,
it is usually altered to talc, carbonate and magnetite, as described above. EMP analyses revealed
that block-like orthopyroxene around olivine is richer in MgO than orthopyroxene in contact to
symplectites and igneous Fe-Ti oxides (En67-71 vs. En58-61).
The symplectite bodies bordering olivine are irregular in shape, with diameters varying
from 0.2 to 5 mm (Fig. 5.5 a-c). They mainly consist of magnetite with ilmenite oxy-exsolution
lamellae set in a polycrystalline orthopyroxene matrix displaying the same composition (En58-61)
as the associated block-like orthopyroxene rim. Within distinct orthopyroxene grains of the
IMOS, two different types of magnetite are intergrown with orthopyroxene, i.e. (i) lamellarplanar, with magnetite grains of 30-60 µm length and 1-7 µm width, and (ii) vermicular, with
magnetite grains of 2-8 µm size (Fig. 5.5 a-f). As shown in perpendicular sections of the same
region, the lamellar-planar and the vermicular types of IMOS are not due to an effect of the third
dimension. In most cases, only one type of intergrowth is present within one individual
orthopyroxene grain, called a symplectitic domain (Fig. 5.5 c). In the lamellar-planar
92
5 Petrography
intergrowths, the magnetite lamellae are commonly curved (Fig. 5.5 a-d). Rarely, lamellar
growth-forms grade into vermicular intergrowths within a single symplectitic domain (Fig. 5.5b).
Fig. 5.5: Polished-section photomicrographs (a-c) and scanning electron microprobe images (e-f) of ilmenitemagnetite-orthopyroxene symplectites, developed in the leucotroctolites of the Kunene Intrusive Complex. Planepolarized light (a-b). Cross-polarized light (c). a) Ilmenite-magnetite-orthopyroxene symplectite in contact to
igneous olivine. Both, the symplectite and olivine are rimmed by a continuous corona of orthopyroxene. b) Within
the whole symplectite body the two types of intergrowths between magnetite and orthopyroxene, i.e. vermicular and
lamellar, are juxtaposed. c) In most cases, only one type of intergrowth occurs within each single orthopyroxene
grain displaying different grey values under half-crossed polarizers. d) One-directional set of ilmenite exsolution
lamellae in lamellar magnetite-orthopyroxene intergrowths. e) Lamellar growthform of magnetite and
orthopyroxene, transitionally grading outward into vermicular intergrowths. f) Anhedral gahnite and ilmenite occur
in patches, with gahnite displaying irregular grain boundaries against magnetite.
5 Petrography
93
The size of the orthopyroxene grains decreases from 4 mm to 0.5 mm when approaching olivine.
An analogous change in grain size is reflected by the Fe-Ti oxides. In direct contact with olivine,
the grain size of the orthopyroxene diminishes to 0.03-1.5 mm, with a simultaneous decrease in
the grain size of the Fe-Ti-oxide lamellae to 3-5 µm length and 0.5-1 µm width (Fig. 5.5 a). As a
result, these intergrowths appear opaque in transmitted light.
The proportions of orthopyroxene and Fe-Ti oxides in the IMOS have been determined
by volumetric integration of the SEM images, utilising a computer-controlled LEICA
microscope with VIDS (video image digitalizing system) software, hosted in the Institute of
Geology, University of Würzburg. It is interesting to note that at the contact to olivine, the finegrained IMOS contains 29-31 (±1) vol.% of Fe-Ti oxide, whereas in the coarsest domains, close
to intercumulus ilmenite, the Fe-Ti oxide content decreases to 13-16 (±1) vol.%. The average
oxide content, integrated across the whole symplectite body is 18 (±1) vol.%. Notwithstanding
the textural differences, compositions of orthopyroxene and magnetite (average composition:
Mt99.1Ulv0.2Chr0.6Glx0.1) remain constant in all symplectite domains. The magnetite itself
displays complex oxy-exsolution of ilmenite (average composition: Ilm93.6Pph4.4Gkl0.9Hem1.1)
lamellae oriented || (111) and || (100). In addition, ilmenite of the same composition occurs in
irregular patches within magnetite (Fig. 5.5 f). Subordinate anhedral gahnite (average
composition: Mt7.5Hc32.3Sp23.1Ga37.1), with grain sizes varying from 1 to 7.5 µm, is associated
with ilmenite in patches, both displaying irregular boundaries against magnetite (Fig. 5.5 f).
SEM studies were undertaken to establish the textural relationships between
orthopyroxene, Fe-Ti oxides and gahnite with the aid of back-scattered electron images (Fig. 5.5
d-f). Depending on the size of the magnetite lamellae, within one single orthopyroxene grain,
two different orientation relationships between magnetite and ilmenite were observed:
(1) Within one symplectite domain, lamellar or vermicular magnetite contains a onedirectional set of parallel ilmenite lamellae oriented parallel to the elongation of the lamellar
magnetite. The orientation of the ilmenite exsolution lamellae is not changed even if the
magnetite host is bent, or by the transition from lamellar to vermicular magnetite (Fig. 5.5 d, e).
Within single magnetite lamellae, the width of the ilmenite oxy-exsolution lamellae is constant
in most cases (Fig. 5.5 d). This feature suggests that individual magnetite lamellae are
monocrystalline, whereas different lamellae represent different crystals. Larger equant magnetite
grains (45-60 µm in size), which occur besides the lamellar magnetite, are presumably
94
5 Petrography
polycrystalline, since the ilmenite lamellae in these grains may be interrupted or changed in
width.
(2) In contrast to (1), one large polycrystalline magnetite aggregate contains ilmenite
lamellae with variable orientation. This may be due to the development of different ilmenite
exsolution-lamellae systems, but may also reflect different orientations of individual magnetite
grains.
In conclusion, the textures suggest an oriented intergrowth of orthopyroxene with
magnetite in a similar way as described by Barton et al. (1991), such that (111) of the magnetite
lamellae is oriented parallel to (100) of orthopyroxene. According to Barton et al. (1991) the
similarity of the symplectites with products of eutectoidal reaction and the absence of hydrous
phases like amphibole or biotite in the symplectites is indicative of nucleation and growth during
subsolidus cooling under static, dry conditions. Small-sized equivalents of the IMOS may occur
within altered olivine crystals, where they nucleated in contact to veins, filled with exsolved
magnetite.
5 Petrography
95
5.2 Petrography of the felsic rock suite
Two samples of quartz-syenite, 4 samples of granite and 29 samples of syenite have been taken
from the sodalite mining area, from a granite intrusion near the southern margin of the Zebra
mountains and along the Kunene River road. It has to be mentioned, however, that more than
half of the 29 syenite samples from the Swartbooisdrif area (N=17) underwent a metasomatic
alteration, and will thus be described later (see Chapter 13.2).
Granite (G):
B-98-170a, B-98-180a, B-99-410, Ku-01-16
Quartz-syenite (S,qtz):
Ku-98-07, Ku-98-40
Syenite (S,fs):
Ku-98-24, Ku-98-93, Ku-98-111, Ku-99-03, Ku-99-11, Ku-99-12,
Ku-99-13, Ku-99-14, Ku-99-16, Ku-99-20, Ku-99-21, Ku-01-13
Both, the alkali-feldspar granite and the quartz-syenite lithologies are characterized by a
equigranular to weakly porphyritic appearance, dominated by medium to fine-grained potassium
feldspar (microcline or orthoclase; 55-75 vol.%) and plagioclase (5-35 vol.%), mostly lacking a
preferred orientation. Dust-like inclusions of hematite-rich Fe-Ti oxides are present in most of
the K-feldspar crystals, accounting for the generally reddish to brownish macroscopic colour of
the rocks. Both the K-feldspar and plagioclase suffered a certain amount of sericitization, which,
however, predominantly affected the K-feldspar and plagioclase cores.
Minor subhedral to euhedral clinopyroxene (3-8 vol.%), locally displaying narrow rims
of euhedral epidote and carbonate, as well as late subhedral to anhedral quartz (5-10 vol.%) are
restricted to the alkali-feldspar granites and the quartz-syenite lithologies. A common feature of
the syenites is the enrichment of Fe-Mg silicates and Fe-Ti oxides, when compared to their
quartz-bearing counterparts. In these rocks subhedral to anhedral hastingsite is the dominant
mineral among Fe-Mg-silicates (5-23 vol.%). The grains of hastingsitic amphibole may in turn
be marginally replaced by symplectitic intergrowths of hastingsite, plagioclase and Fe-Ti oxide.
Common accessories of the felsic rocks are apatite, titanite, epidote, ilmenite, magnetite
and zircon. Due to a brittle deformation event, the feldspar crystals are locally brecciated and
96
5 Petrography
marginally re-crystallized to fine-grained, granular albite mosaics. Cracks in the rocks are
commonly filled by epidote and/or carbonate.
5.2.1 Granite and quartz-syenite
Both the granite (G) and quartz-syenite (S,qtz) samples are mainly composed of micro- to
mesoperthitic potassium feldspar (microcline and occasional orthoclase; Or94-98), containing
albite exsolution lamellae (An1-4). It occurs as medium to coarse-grained, subhedral to anhedral,
tabular grains or laths of up to 1.5 cm in length with embayed to straight grain boundaries. The
K-feldspar most probably crystallized as sanidine which partially to completely retrogressed to
orthoclase and microcline during cooling. Commonly the crystals display Carlsbad twinning and
a strong perthitic exsolution (Fig. 5.6 a, b). String perthite, transgressing into patch and block
perthite is most common (Fig. 5.6 a), whereas other exsolution textures like braid, vein, plate or
flame perthite are rare (Fig. 5.6 b). In places, chess-board albite develops at the expense of
microcline, suggesting an alteration of K-feldspar by auto-metasomatism or fenitization. A
second generation of later microcline may occur, forming separate, anhedral grains, lacking
exsolution textures (Fig. 5.6 c). Both K-feldspar generations are marginally recrystallised to
granular and twinned albite, forming mosaics. Most of the feldspar crystals show a variable
degree of sericitization, commonly concentrated in the crystal core. Exsolved hematite occurs as
very fine-grained inclusions in some of the larger K-feldspar grains or is concentrated along the
grain margins of recrystallised albite.
Weakly reverse zoned, platy plagioclase (An13-18) phenocrysts, with mostly subhedral to
euhedral shape (Fig. 5.6 d) may be present in varying amounts but are always less abundant than
potassium feldspar. The fact that the oligoclase displays straight grain boundaries against the
first generation of potassium feldspar is taken as evidence for a contemporal crystallization of
plagioclase and potassium feldspar, as is typical for subsolvus granites. In one granite sample, an
anti-rapakivi texture is developed, bearing orthoclase with marginal trails of quartz-inclusion that
surrounds heavily sericitized plagioclase cores. Almost pure albite (An1-3) occurs as fine-grained,
granular recrystallisation product of the potassium feldspars and as chess-board albite, replacing
microcline.
Subhedral to anhedral, greenish clinopyroxene (XMg: 0.45-0.51) occurs in the interstices
between feldspar crystals, but may also be present as inclusions in these minerals (Fig 5.6 e). In
5 Petrography
97
Fig. 5.6: Polished-section photomicrographs of the granite and the quartz-syenite. Cross-polarized light (a-e). Planepolarized light (f). a) String perthite grading into blocky patch perthite. b) Vein perthite. c) Late, anhedral
microcline fills the interstices between large K-feldspar crystals. d) Euhedral, tabular plagioclase, bearing an apatite
inclusion, displays straight grain boundaries against K-feldspar of the first generation (upper left). e) Euhedral
clinopyroxene inclusion in plagioclase. f) Anhedral titanite surrounding an euhedral zircon crystal.
98
5 Petrography
cases, the clinopyroxene is replaced by fine-grained intergrowths of epidote, carbonate and
chlorite. These altered clinopyroxenes are locally rimmed by euhedral epidote, which in turn is
surrounded by impersistent rims of granular carbonate.
Anhedral magnetite displaying oxidative exsolution of ilmenite-lamellae ||{111} and
{100} as well as late anhedral quartz (bearing abundant fluid inclusions) occur as separate
grains, filling the interstices between individual feldspar crystals.
Titanite (aSph: 0.87-0.90), as large anhedral grains, is the most common accessory phase,
associated with zircon, epidote or Fe-Ti oxides (Fig. 5.6 f). Apatite occurs as stubby prisms,
commonly included in plagioclase, whereas zircon is found as isolated, euhedral crystals.
Secondary enrichment of epidote (Ps21-32) in clusters is common in samples, where hydrothermal
alteration is more pronounced. Sericite is an alteration product of the feldspars, whereas very
fine-grained epidote, chlorite and carbonate replace former Fe-Mg silicates.
5.2.2 Syenite
Strongly exsolved, micro- to mesoperthitic potassium feldspar (Or91-97) is the main constituent
of the medium-grained syenite. It occurs as subhedral to euhedral, elongated laths of 1-4 mm in
length, commonly displaying Carlsbad twinning (Fig. 5.7 a). Former orthoclase was partially to
completely replaced by microcline during cooling. Common perthite textures are string, patch,
block, plate, braid and hair perthite, whereas flame perthite has only been observed sporadically
(Fig. 5.7 b-d). Like their quartz-bearing counterparts, the syenites contain a second generation of
microcline, forming homogeneous, fine-grained and anhedral crystals. Along their margins and
cracks the potassium feldspars are recrystallised to fine-grained albite mosaics or are replaced by
irregularly bounded aggregates of chess-board albite (Fig. 5.7 f).
Tabular to lathshaped, medium-grained plagioclase is reversely zoned albite to
oligoclase, with the An-contents increasing slightly from core (An9-11) towards the rim (An11-14).
Straight grain boundaries against potassium feldspar are common (Fig. 5.7 e). Plagioclase of
similar composition occurs in fine-grained, symplectitic intergrowths with amphibole and Fe-Ti
oxides. In clear contrast, plagioclase, grown at the expense of former potassium feldspar (Fig.
5.7 f), exhibits a weak normal zonation (An4-5 → An0-2) or is the unzoned, almost pure albite
endmember. Alkali-feldspar inclusions may be preserved in the reversely zoned plagioclase
5 Petrography
99
Fig. 5.7: Polished-section photomicrographs of the syenite. Cross-polarized light (a-f). a) Lathshaped, perthitic
microcline, displaying Carlsbad-twinning. b) K-feldspar braid perthite. c) K-feldspar flame perthite. d) K-feldspar
plate perthite. e) Magmatic plagioclase, displaying straight grain boundaries against K-feldspar. f) Chess-board
albite, replacing K-feldspar.
100
5 Petrography
Fig. 5.8: Polished-section (a-d, f) and ore section (e) photomicrographs of the syenite. Cross-polarized light (f).
Plane-polarized light (a-d). a) Early, euhedral to subhedral clinopyroxene, being completely replaced by fibrous
ferro-hornblende to ferro-actinolite. b) Euhedral hastingsite inclusion in plagioclase. c) Subhedral, interstitial
hastingsite. d) Hastingsite-plagioclase-magnetite-ilmenite symplectite surrounding interstitial hastingsite. e)
Magnetite, displaying curved grain boundaries against ilmenite, is partially replaced by hematite. f) Epidoteankerite-magnetite assemblage, surrounding an altered clinopyroxene.
5 Petrography
101
crystals. In two syenite samples rapakivi-type textures with plagioclase rims on perthitic
orthoclase have been observed. In all other samples, the respective feldspars form discrete
grains.
Texturally early, euhedral to subhedral clinopyroxene is commonly completely replaced
by fibrous ferro-hornblende to ferro-actinolite (XMg: 0.23-0.31; Fig. 5.8 a). Subhedral to
euhedral hastingsite (XMg: 0.01-0.06) is included in plagioclase (Fig. 5.8 b). Hastingsite of a
similar composition may also fill the interstices between individual feldspar crystals (5.8 c). or
form symplectitic intergrowths with plagioclase and Fe-Ti oxides, surrounding interstitial
amphibole (Fig. 5.8 d).
Large, anhedral magnetite is commonly spatially associated with ilmenite, forming
clusters (Fig. 5.8 e). Martitisation of magnetite was observed in various stages of progression,
and may even be advanced to the point where the texture consists of larger hematite areas,
intergrown with relict magnetite. The Fe-Ti oxides are partially surrounded by narrow rims of
titanite. In addition, both ilmenite and magnetite have been observed in very fine-grained
symplectites replacing interstitial hastingsite, where the oxides are accompanied by amphibole
and plagioclase, and as small, isolated grains.
Euhedral, prismatic apatite, anhedral grains and irregular rims of titanite (aSph: 0.790.85) as well as subhedral to euhedral epidote (Ps27-34), intimately associated with magnetite and
ankerite (Fig. 5.8 f), are common accessories (Fig. 5.8 f). Euhedral zircon is less abundant than
in the quartz-syenites and granites. Chlorite (XMg: 0.09-0.11) and sericite, formed at the expense
of hastingsite and feldspar, respectively, are the most common alteration products.
102
6 Mineral chemistry
6 Mineral chemistry
The mineral chemistry of 26 selected samples of the rock types described in Chapter 4.2
(Petrography) has been investigated with the aid of electron-microprobe (EMP) analysis.
White anorthosite (A,w):
Ku-97-08b, Ku-97-33, Ku-97-33a, Ku-97-44
Leucogabbronorite (GN) and pyroxene-bearing anorthosite (A,px):
Ku-97-03, Ku-97-04, Ku-97-13, Ku-97-93, Ku-98-68, Ku-98-71, Ku-98-78, Ku-98-79,
Ku-98-84
Leucotroctolite (T) and olivine-bearing anorthosite (A,ol):
Ku-97-92, Ku-97-95, Ku-97-104, Ku-97-105, Ku-98-52, Ku-98-221a, Ku-98-221b
Quartz-syenite (S,qtz):
Ku-98-07, Ku-98-40
Syenite (S,fs):
Ku-99-11, Ku-99-12, Ku-99-13, Ku-99-14
Representative analyses are compiled in Tables A.5.1.1 to A.5.1.12 of the Appendix A.5.1.
Analytical conditions are provided in Appendix A.2.3.
6.1 Feldspar
Calculation of the feldspar formula is based on 8 oxygens. Representative analyses of feldspar
are depicted in Table A.5.1.1a and b in the Appendix A.5.1.
6.1.1 Anorthositic suite
The vast majority of the analysed cumulus plagioclase of the anorthositic rock suites shows
minor chemical zoning, with the anorthite contents falling in the compositional range of andesine
6 Mineral chemistry
103
to bytownite. In contrast, recrystallised plagioclase mostly exhibits lower An-contents ranging
from An3 to An10. Individual samples generally display only minor variations of the An-contents
of plagioclase. The Or- and Ce-contents of the plagioclase are mostly below 0.5 mol.%.
Relics of unaltered andesine and labradorite (An43-53) in the white anorthosite (Fig. 6.1 a)
mostly exhibit a weak normal zoning with average An-increases of about 2 mol.%. In contrast,
recrystallised plagioclase is inversely zoned albite with An3-4 cores and An5-10 rims.
or
ab
an
50
a) white anorthosite
recrystallised
ab
an
50
b) leucogabbronorite and Px-bearing anorthosite
reversely zoned rims
recrystallised
ab
ab
50
c) leucotroctolite and Ol-bearing anorthosite
50
an
an
d) Gt-bearing leucotroctolite
ab
50
an
Anorthositic suite
white anorthosite
leucogabbronorite
pyroxene-bearing anorthosite
leucotroctolite
olivine-bearing anorthosite
garnet-bearing leucotroctolite
Fig. 6.1: Compositional variability of plagioclase from different rock types of the Kunene Intrusive Complex (white
anorthosite suite and dark anorthosite suite, including leucogabbronorite and pyroxene-bearing anorthosite,
leucotroctolite and olivine-bearing anorthosite, garnet-bearing leucotroctolite) expressed in the diagram orthoclasealbite-anorthite.
104
6 Mineral chemistry
Plagioclase from leucogabbronorites and pyroxene-bearing anorthosites (Fig. 6.1 b) of
the dark anorthosite suite shows oscillatory zoning with the An-contents (An43-54) being similar
to those obtained for the white anorthosites. A tendency of decreasing An-contents towards the
rims is common. A narrow rim of strongly reversed plagioclase was observed in one sample
(Ku-98-84), where the An-content increases abruptly from 55 mol.% to 69 mol.%.
In clear contrast, plagioclase of the olivine-bearing anorthosites and leucotroctolites
(Fig. 6.1 c) shows reverse to oscillatory zoning (An47-75), with the highest An-contents being
displayed by the garnet-bearing leucotroctolite (Fig. 6.1 d) samples. The most extreme
difference between rim and core compositions is 14 mol.%; however, in most of the plagioclases
the An-increase reaches a maximum of 3 mol.%. Plagioclase inclusions in olivine generally
display similar An-contents as the matrix plagioclase. Strongly reversely zoned rims have been
analysed in one sample (Ku-97-95).
In summary, the results indicate an increase of the An-content of plagioclase in the
sequence (1) white anorthosite, leucogabbronorite and pyroxene-bearing anorthosite → (2)
leucotroctolite and olivine-bearing anorthosite → (3) garnet-bearing leucotroctolite, which is in
good accordance with the relative timing of their emplacement. Anderson & Morin (1968)
proposed a subdivision of anorthosite massifs based on the An-contents of plagioclase.
Following their classification, the white anorthosite as well as the leucogabbronorite and
pyroxene-bearing anorthosite of the dark anorthosite suite could be considered as andesine-type
massifs, whereas the younger leucotroctolites and olivine-bearing anorthosites would represent
labradorite-type massifs, even though the compositions of plagioclase overlap. Regarding the
Kunene Intrusive Complex, the increase in the An-content of plagioclase coincides with a
decrease in crustal contamination, as is evidenced by the lack of crustal xenoliths in the
leucotroctolites and olivine-bearing anorthosites. A similar model has also been proposed by
Ashwal (1993) for other anorthosite massifs.
6.1.2 Felsic suite
Potassium feldspar in both, the quartz-syenite and the syenite, exhibits strong perthitic
exsolution, with a potassium feldspar host (Or91-98), containing albite (Ab95-99) exsolution
lamellae (Fig. 6.2). Recalculation of the magmatic compositions of K-feldspar before exsolution
6 Mineral chemistry
105
concentrated on rare crystals, which lack extensive recrystallization and the secondary formation
of chess-board albite.
Reintegrated compositions of K-feldspar cores of the syenite, analysed with a defocused
beam of 20 µm size (Fig. 6.3), fall in the range of Ab43.3An1.9Or
54.8–Ab85.8An4.1Or10.1
whereas
the feldspar rims have a more albite-rich composition (Ab86.8An2.6Or 10.6–Ab88.3An3.1Or8.6) or are
the almost pure albite endmember (Ab95-99).
an
an
an
focused beam
(beam size: 1 µm)
Ku-99-14
or
Ku-98-40
or
Ku-98-07
ab
or
Fig. 6.2: Compositional variability of K-feldspar (mesoperthitic orthoclase/microcline; white symbols) and
plagioclase (black symbols) in the quartz-syenites Ku-98-07 and Ku-98-40 and the syenite Ku-99-14, analysed with
a focused beam of 1 µm size and expressed in the ternary diagram anorthite-albite-orthoclase.
A K-feldspar inclusion, preserved in a reversely zoned plagioclase crystal of the syenite
Ku-99-11, displays an average Or-content of Ab48.5An3.1Or48.4. The quite variable reintegrated
core compositions of the analysed feldspars most probably reflect compositional modifications
during late-stage coarsening of the mesoperthitic texture or continued cation exchange.
Plagioclase of the felsic suite generally displays lower An contents when compared to the
anorthositic rocks. Based on their compositions and zoning patterns, plagioclase may be
subdivided into two groups: Plagioclase of the quartz-syenite (sample Ku-98-40), containing
abundant clinopyroxene inclusions, is almost unzoned oligoclase (An13-18) with a slight rimward
106
6 Mineral chemistry
increase of its An contents (Fig. 6.2). In contrast, plagioclase coexisting with amphibole and/or
surrounding K-feldspar in the syenite (sample Ku-99-11) is mostly albite to oligoclase (Fig. 6.3),
that displays a reverse zonation with the An-contents increasing slightly from core (An9-11) to
rim (An11-14). Recrystallized plagioclase and chess-board albite of both the quartz-syenite (Ku98-07) and the syenite (Ku-99-11, Ku-99-13, Ku-99-14) exhibits a weak normal zonation (An4-5
→ An0-2) or is the almost pure albite endmember (An0-2; Fig. 6.2, 6.3).
an
an
an
defocused beam
(beam size: 20 µm)
Ku-99-13
or
Ku-99-12
or
Ku-99-11
ab
or
Fig. 6.3: Compositional variability of K-feldspar (micro-perthitic orthoclase/microcline, grey symbols) and
plagioclase (black symbols) in the syenite, analysed with a defocused beam of 20 µm size and expressed in the
ternary diagram anorthite-albite-orthoclase.
6.2 Olivine
Microprobe analysis was carried out for olivine from leucotroctolites of the dark anorthosite
suite. Olivine formulae were calculated on the basis of 4 oxygens. Representative analyses are
listed in Table A.5.1.2 in the Appendix A.5.1.
In general the olivines display a restricted, Mg-rich composition, ranging from 54-64
mol.% forsterite (Fo). All olivines are chemically unzoned with respect to the elements analysed.
The Fo-content of olivine versus An-content of plagioclase diagram of Emslie (1985) allows a
6 Mineral chemistry
107
comparison between mafic layered and massif-type intrusions in general. The analysed olivines
of the KIC display only minor compositional variations when compared to the coexisting
plagioclase, the samples thus plot in the field of massif-type anorthosite complexes and show no
mol.% Fo in olivine
affinity to layered intrusions (Fig. 6.4).
90
Dark anorthositic suite
leucotroctolite
80
Ki
Hl
70
St
MM
PI
60
Mi
Ki
Sk
50
40
30
40
50
60
70
80
90
mol.% An in plagioclase
Fig. 6.4: Compositions of coexisting plagioclase and olivine from olivine-bearing anorthosites and leucotroctolites
of the Kunene Intrusive Complex compared to those of other massif-type anorthosites and mafic layered intrusions
(Hl: Harp Lake; Mi: Michikamau; MM: Mealy Mountains; PI: Paul Island. Layered intrusions: Kl: Kiglapait; Sk:
Skaergaard; St: Stillwater ).
The distinctly higher Mg# of olivine when compared to the bulk-rock composition of the
corresponding sample is most probably caused by the additional formation of intercumulus
ilmenite and magnetite. However, the Mg-content of the analysed olivines is strongly controlled
by the composition of the intercumulus melt as is suggested by the positive correlation of
Mg#olivine and Mg#bulk rock.
6.3 Pyroxene
Orthopyroxene has only been analysed in leucotroctolite and white anorthosite samples, since the
analysis of the heavily altered orthopyroxene in the leucogabbronorites and pyroxene-bearing
anorthosites gave no reliable results. Clinopyroxene compositions were determined in
leucogabbronorites and leucotroctolites of the dark anorthosite suite and the felsic suite. The
108
6 Mineral chemistry
calculation of the pyroxene formula followed the method of Papike & Cameron (1980) on the
basis of 8 oxygens (Table A.5.1.3 in the Appendix A.5.1).
6.3.1 Orthopyroxene
Following the classification of Morimoto (1988) all orthopyroxenes of the anorthositic rocks of
the KIC are enstatites, whereas they can be designated as hyperstene and bronzite after
Poldervaart & Hess (1951). The composition of orthopyroxene strongly depends on
the
wo
50
en
a) white anorthosite
fs
interstitial
50
en
b) leucotroctolite
interstitial
en
en
rims around olivine
rims around ilmenite
50
c) Grt-bearing leucotroctolite
interstitial
fs
fs
replacing olivine
50
d) Grt-bearing leucotroctolite
fs
coronae around olivine
en
50
fs
Anorthositic suite
white anorthosite
leucotroctolite
Fig. 6.5: Compositional variability of the composition of orthopyroxene from the white anorthosite and the
leucotroctolite expressed in the diagram wollastonite-enstatite-ferrosilite.
6 Mineral chemistry
109
bordering minerals. Intercumulus grains generally have the highest En-content, which however,
decreases in the sequence white anorthosite → leucotroctolite → Grt-bearing leucotroctolite. In
contrast, the highest Fs-content is displayed by late-magmatic orthopyroxene surrounding
ilmenite and orthopyroxene occurring in symplectitic textures replacing former olivine. Latemagmatic orthopyroxene rims and recrystallised orthopyroxene coronas around olivine display
similar, intermediate compositions, thus suggesting that the recrystallisation was an isochemical
process.
The En-content of orthopyroxene versus An-content of plagioclase diagram of Emslie
(1985) allows a comparison of the analysed orthopyroxene with that of other anorthosite
intrusions. Orthopyroxenes of the KIC display only minor compositional variations when
compared to the coexisting plagioclase, the samples thus clearly do not display a layered
intrusion trend. It has to be mentioned, however, that the En-content of orthopyroxene is also
mol.% En in orthopyroxene
distinctly higher than that of all other yet known massif-type anorthosite complexes.
80
70
B
60
u
St
(GN)
Hl
M
MM
50
L
A
Sk
40
30
Mi
30
40
50
Pl
Du
f
St
(MBZ)
60
70
80
90
mol.% An (Ca/Na) in plagioclase
Anorthositic suite
white anorthosite
leucotroctolite
Fig. 6.6:
Compositions of coexisting plagioclase and orthopyroxene from the white anorthosite and the
leucotroctolites of the Kunene Intrusive Complex compared to those of other massif-type anorthosites and mafic
layered intrusions (A: Adirondack; Hl: Harp Lake; L: Labrieville; M: Morin; Mi: Michikamau; MM: Mealy
Mountains; PI: Paul Island. Layered intrusions: Bu: Bushveld; Duf: Dufek; Sk: Skaergaard; St: Stillwater; MBZ:
middle banded zone).
110
6 Mineral chemistry
A weak positive correlation of the Mg#orthopyroxene and the Mg#bulk
rock
of the
corresponding sample is displayed by most samples of the dark anorthosite suite. However,
orthopyroxene from the white anorthosite does not follow this trend. The most plausible
explanation for this deviation is the low resistance of Mg-Fe silicates against alteration when
compared to Fe-Ti oxides. The hydrothermal overprint of the white anorthosite could thus have
caused a relative decrease in the bulk rock Mg#.
wo
diopside
hedenbergite
augite
en
50
a) leucogabbronorite
fs
interstitial
50
b) leucotroctolite
rims around olivine
rims around ilmenite
50
c) Grt-bearing leucotroctolite
rims around olivine
50
d) quartz-syenite
interstitial
50
Felsic suite
quartz-syenite
Anorthositic suite
leucogabbronorite
leucotroctolite
Fig. 6.7: Compositional variability of clinopyroxene from the white anorthosite and the leucotroctolite of the KIC
and the quartz-syenite of the felsic suite, expressed in the diagram wollastonite-enstatite-ferrosilite.
6 Mineral chemistry
111
6.3.2 Clinopyroxene
Following Poldervaart & Hess (1951) clinopyroxene of the leucogabbronorites is salite
(Wo44.4En32.6Fs18.3–Wo45.7En27.8Fs20.7),
whereas
clinopyroxene
from
the leucotroctolites
(Wo40.03En34.4Fs21.5–Wo43.5En35.9Fs12.4) and garnet-bearing leucotroctolites (Wo42.3En38.3Fs14.2–
Wo45.7En39.1Fs12.3) is salite to augite. After the IMA-classification of Morimoto (1988) all
clinopyroxenes of the anorthositic rocks plot in the fields of diopside and augite (Fig. 6.7). The
Al2O3 content of clinopyroxene from the anorthositic rocks is relatively high, increasing from
leuconorite (Al2O3: 1.49-1.92 wt.%) towards (garnet-bearing) leucotroctolite (Al2O3: 1.65-3.25
wt.%).
The Mg# of clinopyroxene of the anorthositic suite is quite homogeneous but generally
higher than the Mg# of orthopyroxene from similar samples. The Mg#clinopyroxene and the Mg#bulk
rock
exhibit a positive correlation.
In clear contrast, clinopyroxene from the quartz-syenite is more hedenbergitic in
composition (Wo43.8En23.1Fs28.0–Wo45.1En25.4Fs24.3; XMg = 0.45-0.51) when compared to the
anorthositic rocks. The Al contents are generally near the detection limit (Al2O3: 0.41-0.79
wt.%), reflecting the Al-poor whole-rock composition of these rocks.
6.4 Amphibole
Minor amphibole occurs in almost all anorthositic rocks and is the dominant phase among FeMg silicates in the syenites. Amphibole compositions are listed in Table A.5.1.4 in the Appendix
A.5.1. The structural formula of amphibole was calculated on an anhydrous 23-oxygen basis. In
both rock suites the amphiboles can be classified as calcic amphiboles in the sense of Leake et al.
(1997). For the estimation of minimum and maximum Fe3+ contents a calculation modus on the
basis of 15 cations (excluding Na and K) and 13 cations (excluding Na, K and Ca), respectively,
was chosen, following Spear & Kimball (1984). For the calculation of the average Fe3+ contents
the mean value between the two extreme compositions was calculated, following the method of
Papike et al. (1974), modified by Franz & Häussinger (1990).
Following the classification scheme of Leake et al. (1997) the vast majority of the latemagmatic amphiboles from the anorthosite suites, i.e. white anorthosite (Al p.f.u.: 2.42-2.97;
112
6 Mineral chemistry
XMg: 0.65-0.70), leucogabbronorite (Al p.f.u.: 1.96-2.51; XMg: 0.51-0.73) and pyroxene-bearing
anorthosite (Al p.f.u.: 1.82-2.29; XMg: 0.46-0.55) are magnesio-hastingsites/pargasites to ferropargasites/hastingsites (Fig. 6.8). The Ti contents of the amphiboles are variable (Ti p.f.u.: 0.020.40), depending on their textural position, e.g. amphiboles surrounding Fe-Ti oxides are
generally characterized by the highest Ti contents. Magnesio-hastingsites and pargasites from
garnet-bearing corona structures around olivine in the leucotroctolite are characterized by
distinctly higher Mg and Al contents (Al p.f.u.: 2.73-2.81; XMg: 0.77-0.86), which may reflect
c
it i
os
h
t
or nd
an tre
edenite
0.75
magnesiohastingsite
VI
3+
(Al < Fe )
0.50
edenite
0.75
pargasite
VI
3+
(Al > Fe )
0.50
0.25
0.00
1.00
or
t
tre hosi
nd tic
2+
1.00
Ca-amphibole
CaB > 1.50; (Na + K)A > 0.50; Ti < 0.50
an
Ca-amphibole
CaB > 1.50; (Na + K)A > 0.50; Ti < 0.50
Mg/(Mg + Fe )
Mg/(Mg + Fe2+)
the whole rock geochemistry of these rocks.
ferro-pargasite
VI
3+
(Al >Fe )
ferro-edenite
7.0
6.5
6.0
0.25
ferro-edenite
syenite
5.5
Si(T)
0.00
7.0
6.5
6.0
hastingsite
(AlVI < Fe3+ )
5.5
Si(T)
Felsic suite
syenite
Anorthositic suite
white anorthosite
leucogabbronorite
garnet-bearing anorthosite
Fig. 6.8: Compositional ranges of primary magmatic Ca-amphiboles from the anorthositic rocks of the KIC and the
syenites of the felsic suite expressed in the classification diagram of Leake et al. (1997).
In clear contrast, secondary amphibole formed at the expense of clinopyroxene in the
leucogabbronorite and olivine in the leucotroctolites is mostly actinolite and rarely magnesiohornblende (Al p.f.u.: 0.20-0.98; XMg: 0.60-0.65; Fig. 6.9).
Primary magmatic amphibole from the syenites is always hastingsite (Fig. 6.8), which,
however, displays a distinctly lower XMg and Al contents (Al p.f.u.: 1.91-2.10; Ti p.f.u.: 0.070.26; XMg: 0.01-0.06) when compared to late-magmatic amphiboles of the anorthosite suite,
whereas the Ti contents fall in a similar range.
113
2+
Mg/(Mg + Fe )
6 Mineral chemistry
Ca-amphibole
CaB > 1.50; (Na + K)A < 0.50; CaA < 0.50
1.00
tremolite
0.75
actinolite
Felsic suite
syenite
Anorthositic suite
leucogabbronorite
garnet-bearing anorthosite
magnesio-hornblende
anorthositic rocks
0.50
syenite
0.25
ferro-actinolite
0.00
8.0
ferro-hornblende
7.5
7.0
Si(T)
Fig. 6.9: Compositional ranges of secondary Ca-amphiboles from the anorthositic rocks of the KIC and the syenites
of the felsic suite expressed in the classification diagram of Leake et al. (1997).
Secondary amphibole, formed by the hydrothermal alteration of clinopyroxene, is ferrohornblende to ferro-actinolite (Al p.f.u.: 0.22-1.00; XMg: 0.23-0.31; Fig. 6.9). Its higher XMg,
when compared to the magmatic amphiboles of the same samples, is most probably inherited
from the former clinopyroxene.
6.5 Biotite
Microprobe analyses were carried out for biotite from the different anorthositic lithologies.
Biotite formulae were calculated on the basis of 22 oxygens. Representative analyses are
depicted in Table A.5.1.5 in the Appendix A.5.1.
The composition of the analysed biotites is strongly dependent on their respective textural
position and rock type. The classification of Foster (1960; in Tröger, 1982) characterizes the
majority of the unzoned late-magmatic biotite, surrounding ilmenite in the white anorthosite
(XMg: 0.48; AlVI: 0.87; Ti p.f.u.: 0.12), pyroxene-bearing anorthosite (XMg: 0.39-0.46; AlVI:
0.35-0.45;
114
6 Mineral chemistry
3+
VI
4+
R = Al +Ti +Cr
a)
3+
lepidomelane
siderophyllite
siderophyllite
meroxene
(eastonite)
phlogopite
2+
Mg
phlogopite
Fe +Mn
annite
3+
VI
4+
R = Al +Ti +Cr
b)
3+
lepidomelane
siderophyllite
siderophyllite
meroxene
(eastonite)
phlogopite
2+
Fe +Mn
annite
Mg
phlogopite
3+
VI
4+
R = Al +Ti +Cr
c)
3+
lepidomelane
siderophyllite
siderophyllite
meroxene
(eastonite)
phlogopite
2+
Mg
phlogopite
Fe +Mn
annite
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite
Fig. 6.10: Classification of biotite from the anorthositic rocks of the KIC in the modified (Fe2+=Fetot) triangle R3+
(AlVI+Ti+Cr)-Mg-(Fe2++Mn) after Foster (1960, in Tröger, 1982). a) Magmatic biotite from the white anorthosite,
pyroxene-bearing anorthosite and leucogabbronorite. b) Magmatic biotite from the leucotroctolite. c) Secondary
biotite from the garnet-bearing leucotroctolite.
6 Mineral chemistry
115
Ti p.f.u.: 0.25-0.28), leucogabbronorite (XMg: 0.42-0.48; AlVI: 0.43-0.65; Ti p.f.u.: 0.17-0.26)
and garnet-free leucotroctolite (XMg: 0.45-0.48; AlVI: 0.54-0.71; Ti p.f.u.: 0.21-0.26) as
lepidomelanes, whereas meroxenes are rare (Fig. 6.10 a-b). In clear contrast, all biotites formed
by subsolidus processes in the garnet-bearing leucotroctolite samples are meroxenes (Fig. 6.10
c), even though they occur in different textural positions, i.e. as coronas surrounding ilmenite
(XMg: 0.66-0.77; AlVI: 0.20-0.27; Ti p.f.u.: 0.36-0.48) and olivine (XMg: 0.75; AlVI: 0.24; Ti
p.f.u.: 0.34), as fine-grained inclusions in garnet (XMg: 0.68; AlVI: 0.22; Ti p.f.u.: 0.45) and
intergrown with garnet (XMg: 0.0.72-0.73; AlVI: 0.34-0.44; Ti p.f.u.: 0.26-0.33). The fact that
late-magmatic biotite together with amphibole are abundant phases in almost all investigated
samples, suggests that the crystallization from the latest stage liquid differentiates occurred under
a high water fugacity.
6.6 Titanite
Titanite was analysed in samples of the quartz-syenite and the syenite. Titanite formulae were
calculated anhydrous on a 20-oxygens basis (Table A.5.1.6 in the Appendix A.5.1).
Titanite from both the quartz-syenite and the syenite contains appreciable amounts of Al
(Al p.f.u.: 0.28-0.36 and 0.69-0.90, respectively) and Fe3+(Fe p.f.u.: 0.27-0.44 and 0.21-0.43,
respectively), thus deviating significantly from the ideal formula CaTi[O/SiO4]. For the
formation of these so-called grothites, a charge balance by the coupled substitution of (Al, Fe3+)
+ (F, OH-) = Ti4+ + O2- (Oberti et al., 1991) is required, which also accounts for the low total
sums of the analyses.
6.7 Fe-Ti oxides
Ilmenite and magnetite are ubiquitous accessory constituents of both the anorthositic and the
felsic rock suite, whereas martitisation of magnetite is restricted to the syenite samples. The
calculation of the ilmenite and hematite formula was based on 6 oxygens whereas magnetite was
calculated on a 4-oxygens basis. For the estimation of the Fe3+ contents of ilmenite/hematite and
magnetite a calculation modus on the basis of 4 and 3 cations, respectively, was chosen.
116
6 Mineral chemistry
Representative analyses of ilmenite, magnetite and hematite are listed in Tables A.5.1.7, A.5.1.8
and A.5.1.9, respectively, in the Appendix A.5.1.
6.7.1 Ilmenite
In the anorthositic rock suite, ilmenite occurs in three different textural positions, i.e. (1) as
discrete grains in the interstices between plagioclase crystals, (2) as ilmenite exsolution lamellae
in magnetite and (3) in symplectitic intergrowth with orthopyroxene and magnetite. EMP
analyses revealed ilmenite compositions with minor amounts of the Mn end-member pyrophanite
and quite variable contents of hematite. The Mn and Fe3+ contents of intercumulus ilmenite
(MnO: 0.74-3.09 wt.%, Hem: 0.00-4.09 mol.%) slightly increase in the sequence white
anorthosite (MnO: 0.74-2.62), leucogabbronorite and pyroxene-bearing anorthosite (MnO: 1.822.91) towards leucotroctolite and olivine-bearing anorthosite (MnO: 1.28-3.09).
In clear contrast, the MnO contents of ilmenite oxy-exsolution lamellae in magnetite are
distinctly lower, ranging from 0.92-0.99 wt.% (leucotroctolite) to 0.55-1.10 wt.%
(leucogabbronorite). Ilmenite of the ilmenite-magnetite-orthopyroxene symplectites (IMOS) is
characterized by relatively high MnO contents of 1.50-2.09 wt.% in the leucogabbronorites,
1.14-1.24 wt.% in the leucotroctolites and 2.17-3.54 wt.% in the garnet-bearing leucotroctolite
samples. The hematite contents of both ilmenite types fall in a range of 0.00-1.83 mol.%.
Discrete grains of ilmenite in the syenite are characterized by high MnO amounts of 5.376.04 wt.% and a low hematite content of 0.00-2.25 mol.%. In contrast, ilmenite oxy-exsolution
lamellae in magnetite of the quartz-syenite display higher FeOtot/MnO ratios, containing lower
amounts of MnO (4.88-4.94 wt.%) but displaying higher hematite contents (2.39-2.44 mol.%).
6.7.2 Magnetite
Magnetite from the anorthositic rock suite occurs as (1) inclusions in plagioclase, (2)
homogeneous intercumulus grains, (3) intercumulus grains with ilmenite-exsolution lamellae and
(4) in intergrowths with symplectitic ilmenite and orthopyroxene in the IMOS. Depending on
their respective textural position, magnetites display minor differences in their TiO2 contents,
with ranges of 0.59-0.65 wt.% in discrete magnetite grains from the leucotroctolites, 0.03-0.11
wt.% in magnetite displaying ilmenite oxy-exsolution lamellae and 0.02-0.17 wt.% in magnetite
6 Mineral chemistry
117
from symplectitic intergrowths, replacing olivine and pyroxene. Magnetite inclusions in
plagioclase are too small, to obtain reliable analyses.
Magnetite from the felsic suite, displaying ilmenite oxy-exsolution lamellae, is generally
characterized by low Ti contents (TiO2: 0.00-0.13 wt.%). The Ti content of magnetite in
symplectitic intergrowth with amphibole and plagioclase in the syenites is similar (TiO2: 0.02
wt.%).
6.7.3 Hematite
Hematite, formed by martitisation of magnetite from the syenite was found to be almost pure
Fe2O3. The TiO2 contents are generally low, ranging from 0.03-0.18 wt.% whereas Cr2O3, MgO
and MnO are generally below the detection limit.
6.8 Garnet
Andradite-garnet has been observed in one white anorthosite (Ku-97-08b) and one
leucogabbronorite sample (Ku-97-03), where it occurs as xenocrysts that are marginally replaced
by chlorite. In addition, almandine-pyrope garnet surrounding intercumulus olivine and ilmenite
is present in some of the leucotroctolite samples (Ku-97-104, Ku-97-105, Ku-98-57, Ku-98-220,
Ku-98-221a, Ku-98-221b). In contrast to the former, these garnets were formed during a
subsolidus re-equilibration of the leucotroctolites under upper amphibolite to granulite-facies
conditions. Garnet formulae were calculated on a 24-oxygen basis. For the estimation of the Fe3+
contents of garnet a calculation modus on the basis of 16 cations was chosen. End-member
calculation followed the sequence uvarovite, andradite, grossular, almandine, spessartine and
pyrope. Representative analyses are depicted in Table A.5.1.10 in the Appendix A.5.1.
Garnet xenocrysts in the white anorthosite and leucogabbronorite belong to the andraditegrossular solid-solution series, ranging between And72.3Grs26.9 in sample Ku-97-08b and
And94.9Grs4.4 in the leucogabbronorite sample Ku-97-03 and containing negligible amounts of Ti,
Mg and Mn. The compositional range of the analysed garnets agrees well with that known from
typical calc-silicate rocks and skarn assemblages. Indeed, scattered outcrops of calc-silicate
rocks have been observed in both the Namibian (Brandt, in prep.) and the Angolan part of the
Epupa Complex (Carvalho & Alves, 1990), thus suggesting that the early generations of
118
6 Mineral chemistry
anorthositic crystal mushes incorporated crustal calc-silicate material during their ascent.
However, such calc-silicate rocks are rare in the exposed parts of the Epupa Complex, which
mainly consist of metabasites, metapelites, metasemipelites and granitic gneisses (Brandt et al.,
2003). The fact that xenoliths of the latter lithologies have never been observed in the
anorthositic rocks is taken as evidence that the lower, not exposed zones of the EC differ in
composition from its superficial parts by containing significant amounts of calc-silicate material.
Garnet of the leucotroctolites is always almandine-pyrope with a high grossular
component (XGrs: 0.04-0.18), reflecting the Ca-rich whole-rock geochemistry of the
leucotroctolites. Spessartine is generally low (XSps: 0.02-0.04) and unzoned, indicating
intracrystalline diffusional exchange and/or rapid growth. Depending on their mineral
assemblage the analysed garnets display significant compositional differences (Fig. 6.11):
0.9
a
c
b
d
0.8
orthopyroxene
crack
crack
crack
orthopyroxene
orthopyroxene
crack
biotite
plagioclase
biotite
plagioclase
biotite inclusion
biotite inclusion
0.3
biotite inclusion
0.4
ilmenite
0.5
biotite inclusion
0.6
biotite inclusion
XAlm, XPrp, XGrs, XSps, XFe
0.7
0.2
0.1
0.0
0
100
200
300
0
20
35
0
20
40
60
80
0
20
40
60
garnet size (µm)
Fig. 6.11: Zoning profiles of garnet from the garnet-bearing leucotroctolites of the KIC. a) Poikiloblastic garnet
from a garnet-biotite corona surrounding ilmenite in sample Ku-98-221b. b) Inclusion-free garnet from a garnetbiotite corona surrounding ilmenite in sample Ku-98-221b. c) Garnet from a garnet-orthopyroxene-quartz corona
surrounding olivine in sample Ku-98-221b. d) Garnet from a garnet-orthopyroxene-quartz corona surrounding
olivine in sample Ku-98-221a.
•
Garnet in corona textures surrounding olivine has an average composition of
Prp28.6Sps3.4Alm54.9Grs8.6And4.5 (XMg: 0.31-0.42; Fig. 6.11 c-d). When bordered by
plagioclase and orthopyroxene garnets from type I coronas exhibit a weakly developed
asymmetric zonation, involving slight decreases in XPrp and XGrs and increases in XAlm and
XFe with increasing proximity to orthopyroxene (Fig. 6.11 c). The highest pyrope contents
are preserved in narrow zones, which presumably remained unaffected during the
retrograde re-equilibration.
XFe
XAlm
XPrp
XGrs
XSps
6 Mineral chemistry
•
119
When compared to garnet surrounding olivine, garnet from type II corona textures around
ilmenite is characterized by distinctly lower pyrope contents (average composition:
Prp22.7Sps3.3Alm59.2Grs10.6And4.2; XMg: 0.23-0.30; Fig. 6.11 a-b) and lacks significant
zoning patterns. Lowest pyrope contents were obtained for a poikiloblastic garnet, bearing
abundant biotite inclusions (Fig. 6.11 a), whereas inclusion-free garnet generally exhibits
significantly higher pyrope contents (Fig. 6.11 b). It can thus be concluded, that
poikiloblastic garnet was affected by retrograde Fe-Mg exchange with the included biotite,
resulting in an almost complete homogenization of the garnet host, whereas the XAlm and
XPrp of inclusion-free garnet remained almost unchanged.
6.9 Epidote
Accessory epidote minerals occur in almost all rock samples of both the anorthositic and felsic
suite as a retrograde phenomenon. Epidote formulae were calculated on an anhydrous basis of 25
oxygens. Representative analyses are compiled in Table A.5.1.11 in the Appendix A.5.1.
Compositional variations can be expressed in terms of the pistacite (Ps) component of epidote
(mole fraction of Ca2Fe3Si3O12(OH)). Significant core-rim relationships of the epidotes do not
occur.
In the anorthositic rock suite, the pistacite component of epidote varies between 0.14-0.28
mol.%, corresponding to Al-epidotes in the sense of Holdaway (1972), whereas epidotes of the
quartz-syenites (Ps: 0.21-0.32 mol.%) and syenites (Ps: 0.27-0.34 mol.%) are generally more Ferich and can thus be described as Fe-epidotes.
6.10 Chlorite
Chlorite in both the anorthositic and felsic suite is formed by the hydrothermal alteration of FeMg silicates (biotite and pyroxene in the anorthositic suite and amphibole in the syenites).
Calculation of the formula is based on 28 oxygens, excluding H2O. The estimation of Fe3+
followed the method of Laird & Albee (1981). Due to the small grain size of chlorite and its
intimate association with biotite, amphibole, carbonate, epidote or actinolite only a few reliable
analyses were obtained (Table A.5.1.12 in the Appendix A.5.1).
120
6 Mineral chemistry
1.0
daphnite
0.8
brunsvigite
pseudothuringite
ripidolite
0.6
diabantite
0.4
pyknochlorite
corundophilite
Felsic suite
syenite
0.2
Anorthositic suite
leucogabbronorite sheridanite
leucotroctolite
0.0
4.4
5.0
clinochlore
5.6
penninite
6.2
6.8
Si (T)
Fig. 6.12: Classification of chlorite from the anorthositic rocks of the KIC and the syenites of the felsic suite in the
classification diagram of Hey (1954).
Following the classification of Hey (1954), chlorites formed by the alteration of biotite
from the anorthositic rock suite is pyknochlorite, whereas those growing at the expense of
pyroxene are ripidolite (Fig. 6.12). In clear contrast, the chlorites in the syenites can be classified
as Fe-rich brunsvigite (Fig. 6.12).
6.11 Comparative approach
In summary, the respective mineral compositions of the granitoid suite appear to extend the
compositional range of the anorthositic rocks towards higher Fe/Mg ratios and lower Al of the
Fe-Mg silicates (clinopyroxene, orthopyroxene, amphibole, epidote and chlorite) and lower An
contents of plagioclase. The relative changes in the mineral compositions could provide evidence
for a consanguinity between the two rock suites.
However, intermediate mineral compositions in between those of the anorthositic and the
felsic suite are missing. Moreover, the relative changes of the mineral chemistry, displayed by
the subsequently emplaced anorthosite lithologies themselves rather imply an opposite
6 Mineral chemistry
121
evolutionary trend, including an increase of the XMg of amphibole and the An-content of
plagioclase with subsequent intrusion. Based on the mineral chemistry, the two rock suites are
thus interpreted to represent chemically independent suites.
122
7 Geochemistry
7 Geochemistry
7.1 Major and trace element geochemistry
Geochemical data was obtained for 56 rock samples of the two anorthosite suites and 12 rock
samples of the felsic suite. The investigation of the bulk rock geochemistry shall serve as a tool
for characterising the nature and magmatic evolution of the basic and the felsic melts. Thus,
samples affected by secondary metasomatic alteration were excluded from this part of the study
but will be discussed later. The obtained major and trace element chemical data are given in
Table A.5.4.1 in the Appendix A.5.4. Analytical conditions are given in the Appendix A.2.4.
Felsic suite
Granite (G): B-98-170a, B-98-180a, B-99-410
Quartz-syenite (S,qtz): Ku-98-40
Syenite (S,fs): Ku-98-93, Ku-99-03, Ku-99-11, Ku-99-12, Ku-99-13, Ku-99-14, Ku-99-16, Ku-99-20
White anorthosite suite (A,w)
Ku-97-08b, Ku-97-33, Ku-97-33aw and Ku-97-33ag (i.e. the white and the green portions of sample
Ku-97-33a), Ku-97-33b, Ku-97-44, Ku-97-99a, Ku-98-45, Ku-98-60
Dark anorthosite suite
Pyroxene-bearing anorthosite (A,px): Ku-97-02, Ku-97-12, Ku-97-13, Ku-97-31, Ku-97-92a, Ku97-93, Ku-98-123
Leucogabbronorite (GN): Ku-97-03, Ku-97-03b, Ku-97-03c, Ku-97-04, Ku-97-05, Ku-97-06, Ku97-08a, Ku-97-30, Ku-97-31b, Ku-98-68, Ku-98-92, Ku-98-126, Ku-98-128, Ku-98-129, Ku98-222, B-98-101a
Olivine-bearing anorthosite (A,ol): Ku-97-28, Ku-97-28a, Ku-97-95, Ku-98-23, Ku-98-124, B-98274a, B-98-274b, B-98-281b
Leucotroctolite (T): Ku-97-92, Ku-97-96, Ku-98-41, Ku-98-52, Ku-98-53, Ku-98-94, Ku-98-121,
Ku-98-125, Ku-98-228, Ku-98-230, Ku-98-231
including the garnet-bearing samples (T,grt): Ku-97-104, Ku-97-105, Ku-98-220, Ku-98221a, Ku-98-221b
7.1.1 Anorthositic suite
Following the classification of Tischendorf (in Seim & Tischendorf, 1990) most unaltered
samples of the dark anorthosite suite indeed fall into the compositional range of anorthosites
7 Geochemistry
123
(Fig. 7.1). However, the pervasively altered white anorthosite samples strongly deviate from the
average anorthosite composition due to the use of Na2O and K2O as discrimination factors,
which were strongly enriched by secondary processes affecting these rocks.
22
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite
Na2O +
K2O
(wt.%)
20
18
urtite
16
foyaite
naujaite
NE-SYENITE
14
miaskite
alkalisyenite
12
nordmarkite
ijolite
orthoclasite
quartz-syenite
10
THERALITE
FOIDOLITE
8
6
SYENITE
KFS-GRANITE
syeno-granite
MONZONITE / QZ-MONZONITE
GRANODIORITE / MONZOGRANITE
missourite
melteigite
turjanite
monzodiorite / qz-monzodiorite
DIORITE / QZ-DIORITE
4
MELILITOLITE
2
kurdite
trondhjemite
TONALITE
plagiogranite
gabbro
ferrodiorite
jakupirangite
uncompahgrite
rapakivi-granite
shonkinite
essexite
teschenite
okaite
GABBRONORITE
ANORTHOSITE
hornblendite
wehrlite
LHERZOLITE troctolite PYROXENITE
harzburgite
websterite
dunite
PERIDOTITE
0
30
34
alkali-granite
38
42
46
ultrabasic
50
basic
SiO2 (wt.%)
54
58
62
intermediate
66
70
74
78
acidic
Fig. 7.1: Compositional variability of the anorthositic rocks of the KIC expressed in the classification diagram of
Tischendorf (in Seim & Tischendorf, 1990).
In order to constrain the particular style of hydrothermal alteration, which affected the
white anorthosite, the graphical representation of alteration indexes ("alteration box plot") by
Large et al. (2001) is used (Fig. 7.2). In clear contrast to the least altered samples of the dark
anorthosite suite, white anorthosite samples mainly follow a trend of albitization/chloritization,
which is also displayed by altered leucogabbronorite samples of the dark anorthosite suite.
Both the white and the dark anorthosite suite are enriched in Al2O3 (21-29 and 16-30
wt.%), CaO (4-13 and 7-13 wt.%), Na2O (4-9 and 3-6 wt.%) and SiO2 (45-51 and 38-54 wt.%),
reflecting high amounts of cumulus plagioclase, whereas they are deficient in all other oxides.
CaO/Na2O is quite variable, but weakly increases in the sequence white anorthosite and
124
7 Geochemistry
leucogabbronorite towards leucotroctolite, reflecting differences in their respective plagioclase
compositions.
100
epidote/calcite
chlorite
least altered
anorthositic rocks
90
Felsic suite
granite
quartz-syenite
syenite
80
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite
60
itis
atio
n
50
40
alb
CCP Inde x
70
30
felsic rock suite
sericite
20
10
0
albite
0
K-feldspar
10
20
30
40
50
60
Alteration Index
70
80
90
100
Fig. 7.2: Alteration styles of the white anorthosite suite and altered samples of the dark anorthosite suite of the
Kunene Intrusive Complex, expressed in the modified alteration box plot of Large et al. (2001). Compositional
ranges of the granite, quartz-syenite and syenite of the felsic suite are given for comparison. (CCP Index:
CCPI =
100 × ( MgO + FeO)
( MgO + FeO + Na2O + K 2O)
, Alteration Index: AI =
100 × ( K 2O + MgO)
( K 2O + MgO + Na2O + CaO)
).
The Fetot/(Fetot+Mg) ratio is generally intermediate (> 0.28), pointing to anorthosite
crystallization from fractionated melts. However, the highest values are displayed by the white
anorthosite (0.51-0.74) and the leucogabbronorite (0.31-0.66), which both were affected by
crustal contamination. The geochemical trends of the major elements of least altered samples of
the dark anorthosite suite, as demonstrated in various Harker diagrams (Fig. 7.3), correlate well
with varying proportions of cumulus plagioclase and intercumulus phases (i.e. Fe-Mg silicates
and Fe-Ti oxides). Decreases of Si, Al, Ca and Na are accompanied by increases of Fe, Mn and
Mg from anorthosite and troctolite towards leucogabbronorite of the dark anorthosite suite.
Leucogabbronorites generally display the highest values of K and P, reflecting variable
but generally high amounts of late-magmatic biotite and apatite, which crystallized from pockets
of intercumulus melt. Trends of the white anorthosite are mostly indistinguishable from those
of the
7 Geochemistry
125
14.0
4.5
4.0
Felsic suite
granite
quartz-syenite
syenite
3.5
10.0
2.5
2.0
CaO (wt.% )
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite
3.0
TiO2 (wt.% )
12.0
1.5
8.0
6.0
4.0
1.0
2.0
0.5
0.0
40
45
50
55
60
65
70
75
0.0
80
40
45
50
55
34.0
Na2O (wt.% )
Al 2O3 (wt.% )
24.0
22.0
20.0
18.0
70
75
80
65
70
75
80
65
70
75
5.0
4.0
2.0
14.0
1.0
12.0
45
50
55
60
65
70
75
0.0
80
40
45
50
55
SiO 2 (wt.%)
12.0
MgO (wt.% )
10.0
8.0
6.0
4.0
2.0
45
50
55
60
65
70
75
80
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
40
45
50
55
SiO 2 (wt.%)
8.0
7.0
6.0
V (ppm)
5.0
4.0
3.0
2.0
1.0
50
55
60
60
SiO 2(wt.%)
9.0
45
60
SiO 2 (wt.%)
14.0
FeOt (wt.% )
65
6.0
3.0
16.0
K2 O (wt.% )
80
7.0
26.0
0.0
40
75
8.0
28.0
40
70
9.0
30.0
0.0
65
10.0
32.0
10.0
40
60
SiO 2 (wt.%)
SiO2 (wt.%)
65
70
75
80
SiO 2 (wt.%)
300
280
260
240
220
200
180
160
140
120
100
80
60
40
20
0
40
45
50
55
60
80
SiO 2 (wt.%)
Fig. 7.3: Variation of major element contents of the anorthositic rocks of the Kunene Intrusive Complex and the
granites, quartz-syenites and syenites of the felsic suite.
126
7 Geochemistry
dark anorthosite suite, with the exception of Na and K, that were strongly enriched during the
hydrothermal alteration.
Al2O3
An
FeO+Fe2O3
mo
l.%
40
50
60
Harp
Lake
Morin
Harp Lake
Morin
MgO
Na2O+K2O
MgO+FeO
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite
Opx, Ol
MgO+FeOt
CaO
t
Harp Lake
Morin
Na2O
40
50
60 70 80 90
mol.% An
CaO
Fig. 7.4: Composition of the anorthositic rocks of the KIC, expressed in triangular diagrams (AFM-, ACF- and
FNC-diagrams). Geochemical data for the Harp Lake massif, Labrador (Emslie, 1980) and the Morin massif,
Quebec (Martignole, 1974; Emslie, 1975b) was plotted for comparison.
In order to compare the major element geochemistry of the KIC with metamorphosed and
not metamorphosed anorthosite massifs, a set of triangular diagrams (AFM, FNC and ACF
diagrams; Fig. 7.4) was used. The anorthositic rock samples of the KIC are characterized by a
wide range of rock compositions, which, however, display a certain continuity throughout all
diagrams. The scattering of data points in the ACF and FNC diagrams illustrate the variability of
plagioclase compositions of the white and the dark anorthosite suite, whereas the trends towards
higher FeO+MgO are reflected by increasing amounts of intercumulus silicates and oxides,
which reach highest contents in the leucogabbronorites and leucotroctolites. Trends of the white
7 Geochemistry
127
anorthosite and the dark anorthosite suite are nearly indistinguishable, thus suggesting that the
older white anorthosite suite was composed of a variety of anorthositic lithologies before its
pervasive alteration. The obtained compositional ranges of the dark anorthosite suite correlate
best with those obtained for the Harp Lake massif, Labrador, whereas they deviate from those of
the metamorphosed Morin massif, Quebec (Fig. 7.4).
The highest values for Ni are obtained for the olivine-bearing anorthosites and
leucotroctolites, corresponding to increasing modal amounts olivine in these rocks. In contrast,
the high V and Zn contents of the leucogabbronorites are reflected by higher modal amounts of
intercumulus ilmenite and pyroxene.
1000
a)
1000
white anorthosite
100
rock / chondrite
rock / chondrite
100
10
1
10
pyroxene-bearing
anorthosite
0
Ba Rb Th K Nb Ta La Ce Sr Nd P Sm Zr Hf Ti Tb Y TmYb
c)
1000
olivine-bearing
anorthosite
100
d)
leucotroctolite
100
rock / chondrite
rock / chondrite
leucogabbronorite
1
0
1000
b)
10
1
10
1
0
0
Fig. 7.5: Chondrite-normalized (except K, Rb and P) trace element abundance diagrams (spider-grams) for the
white anorthosite suite and the dark anorthosite suite of the Kunene Intrusive Complex, using the normalization
factors of Thompson et al. (1983; see Thompson, 1982, for details on the selection of normalization factors for K,
Rb and P).
Chondrite normalized trace element plots of all rock samples investigated show a strong negative
slope, with enrichment in Ba, Rb, K and Sr relative to Tb, Y, Tm and Yb (Fig. 7.5). A variably
128
7 Geochemistry
accentuated positive Sr anomaly is displayed by all lithologies, pointing to a Sr-enriched parental
melt and thus ruling out extensive plagioclase crystallization prior to the emplacement of the
anorthosite massifs. The highest Sr-concentrations have been obtained for the white anorthosite
samples
(540-1250
ppm; Fig. 7.5 a) and the pyroxene-bearing anorthosites and
leucogabbronorites (380-980 ppm; Fig. 7.5 b) of the dark anorthosite suite, which, at the same
time, exhibit the highest overall trace element abundances and the lowest An contents of
plagioclase. In contrast, the lowest average Sr concentrations (470-640 ppm) have been obtained
for the youngest intrusives, the olivine-bearing anorthosites (Fig. 7.5 c) and leucotroctolites (Fig.
7.5 d), corresponding with the highest average An-contents of plagioclase and low total sums of
incompatible trace elements. An inverse correlation between Sr and the An-content is a common
feature of layered intrusions and is commonly attributed to plagioclase fractionation, causing a
decrease of the Sr content in the melt (see Ashwal, 1993, for a review). However, in the case of
the Kunene Intrusive Complex, this trend is rather caused by decreasing degrees of crustal
contamination during the subsequent intrusion of the anorthosite massifs. This interpretation is
constrained by comparably high values of K, Rb and Ba of both the white anorthosites and the
leucogabbronorite samples. However, in case of the white anorthosite, these trends may also be
influenced by a hydrothermal alteration.
7.1.2 Felsic suite
According to the classification of Tischendorf (in Seim & Tischendorf, 1990) the felsic suite can
be subdivided in K-feldspar granites, quartz-syenites and syenites (Fig. 7.6). The SiO2 contents
of the granites range between 71-76 wt.%, whereas those of the quartz-syenite (SiO2: 66 wt.%)
and the syenites (SiO2: 58-62 wt.%) are distinctly lower (Fig. 7.3). In clear contrast to the
anorthositic rock suite, all felsic samples are enriched in K2O (2.4-8.7 wt.%) and Na2O (3.1-8.3
wt.%) relative to CaO (0.3-4.4 wt.%). The data for samples of the felsic rock suite form coherent
arrays on major element variation diagrams, involving a negative correlation of Si with Ti, Al,
Fe, Mn, Mg, Ca, Na and P whereas K increases in the same direction (Fig. 7.3). In case of Al, Ca
and Na these trends are unequal to those obtained for the anorthositic rocks and thus strongly
argue against a chemical relation between the two suites. The high scatter of data points for Al,
Na and K are reflected by local variations in the abundances of K-feldspar and plagioclase. The
ratio FeOtot/(MgO+FeOtot) is constantly high, ranging between 0.91-0.96 in the granite and
quartz-syenite and 0.85-0.98 in the syenite and thus plotting in the ferroan field of Frost et al.
(2002; Fig. 7.7 a). Syenites and the quartz-syenite are alkalic, whereas the granites are alkali-
7 Geochemistry
129
calcic (Fig. 7.7 b). Most of the felsic samples analysed are metaluminous to slightly
peraluminous with increasing SiO2, except for one peraluminous syenite sample (Ku-99-16; Fig.
7.7c). However, none of the samples is peralkaline, as would be typical for mantle-derived melts.
22
Na2O +
K2O
(wt.%)
20
18
Felsic suite
granite
quartz-syenite
syenite
urtite
16
foyaite
naujaite
NE-SYENITE
14
miaskite
alkalisyenite
12
nordmarkite
ijolite
orthoclasite
quartz-syenite
10
THERALITE
FOIDOLITE
8
6
SYENITE
KFS-GRANITE
syeno-granite
MONZONITE / QZ-MONZONITE
GRANODIORITE / MONZOGRANITE
missourite
melteigite
turjanite
monzodiorite / qz-monzodiorite
DIORITE / QZ-DIORITE
4
MELILITOLITE
2
kurdite
rapakivi-granite
shonkinite
essexite
teschenite
okaite
alkali-granite
trondhjemite
TONALITE
plagiogranite
gabbro
ferrodiorite
ANORTHOSITE
jakupirangite
uncompahgrite
GABBRONORITE
hornblendite
wehrlite
LHERZOLITE troctolite PYROXENITE
websterite
PERIDOTITE
SiO2 (wt.%)
harzburgite
dunite
0
30
34
38
ultrabasic
42
46
50
basic
54
58
62
intermediate
66
70
74
78
acidic
Fig. 7.6: Compositional variability of the felsic rock suite (granite, quartz-syenite and syenite), expressed in the
classification diagram of Tischendorf (in Seim & Tischendorf, 1990).
All samples contain abundant LIL (Rb, Ba, REE) and HFS elements (Zr, Y, Nb), typical
of anorogenic (A-type) granites (Fig. 7.7 d). In the discrimination diagrams Y against Nb and
(Y+Nb) against Rb of Pearce et al. (1984) all samples plot in the field of within-plate granites
(WPG; Fig. 7.7 e-f), broadly following a continental rift trend as defined by Pearce et al. (1984).
Chondrite normalized trace element plots show depletion in Nb, Sr, P and Ti and a strong
negative slope, except the accentuated positive Zr anomaly (Fig. 7.8). Remarkably, the traceelement patterns of the quartz-syenite and the syenite are quite similar to those of common upper
continental crust rocks as granites or metapelites (see Thompson et al., 1983, for a review), thus
suggesting that the felsic rocks have evolved from mixtures of crustal and mantle sources.
130
7 Geochemistry
12
a)
0.9
Na2O+K2O -CaO
8
0.8
t
ferroan n
sia
magne
0.7
Felsic suite
granite
quartz-syenite
syenite
0.6
0.5
45
50
55
60
65
SiO2 (wt.%)
70
75
alkalic
6
cic
lic
cal
lka
i
a
l
a
c
alk
cal
calcic
4
2
0
-2
-4
-6
-8
80
45
100
2.8
c)
2.4
meta-aluminous
50
55
60
65
SiO2 (wt.%)
2.0
1.6
70
75
80
d)
peraluminous
A-type granite
FeOtot/MgO
Al2O3/(Na2O+K2O) molar
b)
10
t
FeO /(FeO +MgO)
1.0
Fractionated
granites
10
1.2
Non-fractionated
I- and S-type
granites
0.8
0.4
peralkaline
0.6
0.8
1.0
1.2
ASI
1.4
1.6
1.8
1
2.0
10
1000
e)
100
1000
Zr+Nb+Y+Ce (ppm)
f)
10000
syn-COLG
WPG
100
10
10
VAG +
syn-COLG
1
WPG
Rb (ppm)
Nb (ppm)
100
VAG
ORG
1
1
10
100
Y (ppm)
1
ORG
10
100
1000
Y+Nb (ppm)
Fig. 7.7: Compositional characteristics of the felsic suite expressed in various binary discrimination diagrams. a)
Plot of FeOtot/(FeOtot+MgO) against SiO2 with the boundary between ferroan and magnesian fields given by Frost et
al. (2002). b) Plot of the modified alkali-lime index Na2O+K2O-CaO against SiO2 (Frost et al., 2002). c) Shand´s
alkalinity index (plot of Al2O3/(Na2O+K2O)molar against Al2O3/(Na2O+K2O+CaO)molar). d) A-type granitoid
discrimination diagram after Whalen et al. (1987). e) and f) Tectonic environment discrimination diagrams after
Pearce et al. (1984).
7 Geochemistry
131
1000
quartz-syenite
syenite
rock / chondrite
100
10
1
anorthositic rock suite
0
Fig. 7.8: Chondrite-normalized (except Rb, K, P) trace element abundance diagrams (spider-grams) for the syenites
of the felsic suite. Compositional ranges of the anorthositic rocks of the Kunene Intrusive Complex (light grey) and
representative fusible rock types of the continental crust (Thompson et al., 1983; dark grey) are marked for
comparison.
7.2 REE geochemistry
REE-analyses were carried out on fourteen samples of the anorthositic suite and on two samples
of the felsic suite. Analytical results are given in Table A.5.4.2 in the Appendix A.5.4.
Regarding their rare earth element (REE) patterns, all anorthositic rocks are
characterized by a variable enrichment of the light REE (LaN/YbN: 3.3-26.9), a strong positive
Eu (Eu/Eu*: 1.3-11.2) anomaly and flattening in the heavy REE (Fig. 7.9). In the dark
anorthosite, the overall abundance of REE increases in the sequence leucotroctolite (Eu/Eu*:
5.1-11.2; LaN/YbN: 10.1-26.9) – olivine-bearing anorthosite (Eu/Eu*: 4.4-5.2; LaN/YbN: 12.816.3) - pyroxene-bearing anorthosite (Eu/Eu*: 3.0-4.0¸ LaN/YbN: 8.9-9.9) – leucogabbronorite
(Eu/Eu*: 1.3-1.8¸ LaN/YbN: 3.3-6.3), resulting in a relative decrease of the Eu anomaly. The
white anorthosite shows strongly varying REE patterns, one (Ku-97-45; Eu/Eu*: 7.1¸ LaN/YbN:
26.9) resembling that of the leucotroctolites of the dark anorthosite suite whereas two other
patterns (Ku-98-44, Ku-97-33b; Eu/Eu*: 1.6-2.3; LaN/YbN: 5.9-12.6) agree well with those of
the leucogabbronorites and pyroxene-bearing anorthosites. It may thus be concluded that the
white anorthosite, like the dark anorthosite, represents a suite of different anorthositic lithologies
rather than an anorthosite sensu stricto as was proposed by Menge (1998). When compared to
the corresponding lithologies of the dark anorthosite suite, all three white anorthosite samples
display a slight enrichment in light the REE and a relative decrease of the heavy REE as
132
7 Geochemistry
evidenced by their LaN/YbN ratios. This steeping of their REE patterns supports evidence for an
elevated amount of crustal contamination and/or the pervasive alteration during their intrusion.
100
100
pyroxene-bearing
anorthosite
10
olivine-bearing
anorthosite
1
0.1
leucotroctolite
rock / chondrite
rock / chondrite
anorthosite
a)
La Ce Pr
10
1
0.1
Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
b)
La Ce Pr
100
100
white anorthosite
10
rock / chondrite
rock / chondrite
leucogabbronorite
1
0.1
c)
La Ce Pr
10
1
0.1
d)
La Ce Pr
Fig. 7.9: Diagrams showing the variations in chondrite-normalized rare earth element (REE) abundances for the
anorthositic rocks of the KIC. a) Olivine-bearing anorthosite (Ku-97-28, Ku-98-23) and pyroxene-bearing
anorthosite (Ku-97-13, Ku-97-31). b) leucotroctolite (Ku-97-92, Ku-97-104, Ku-98-41, Ku-98-230). c)
leucogabbronorite (Ku-97-04, Ku-97-05, Ku-97-31b, Ku-98-68). d) white anorthosite (Ku-97-33b, Ku-97-44, Ku98-45).
Chondrite-normalized REE plots of the quartz-syenite and the syenite are almost
identical. Low LaN/YbN ratios of 5.9-9.5 indicate a moderate LREE enrichment (Fig. 7.10). With
Eu/Eu* ratios varying from 1.01 to 1.06, the rocks lack an accentuated europium anomaly. An
extensive plagioclase accumulation prior to their emplacement, as evidenced by an accentuated
Sr trough in the chondrite normalized trace element plots, can thus be ruled out.
7 Geochemistry
133
rock / chondrite
1000
100
quartz-syenite
syenite
anorth
ositic r
ock su
ite
10
1
0.1
La Ce Pr
Fig. 7.10: Diagrams showing the variations in chondrite-normalized REE abundances for the syenites of the felsic
suite compared to those of the anorthositic rocks of the KIC (quartz-syenite Ku-98-40, syenite Ku-99-20).
The results obtained strongly suggest that the anorthositic and the felsic lithologies
represent two chemically independent suites, with the melts parental to the anorthositic rocks
being mantle-derived liquids, that were modified by only minor crustal contamination during
their uprise, whereas the felsic melts parental to the granitoids are suggested to have been
generated in the crust.
1.0
15
0.9
Na2O + K2O (wt.%)
Fe*/(Fe*+Mg)
0.8
0.7
0.6
Felsic suite
granite
quartz-syenite
syenite
0.5
0.4
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite
0.3
0.2
0.1
0.0
20
30
40
50
SiO2 (wt.%)
10
alkalic
5
subalkalic
0
60
70
80
20
30
40
50
60
70
80
SiO2 (wt.%)
Fig. 7.11: Compositional variability of both the anorthositic and the felsic rock suite compared to similar lithologies
of the Harp Lake massif, Labrador (stippled boundaries; Emslie, 1980) a) Plot of molar Fetot/(Fetot+Mg) against SiO2
and b) Na2O+K2O against SiO2.
134
7 Geochemistry
Like the Harp Lake massif, Labrador, the two Namibian rock suites are characterized by
chemical discontinuities, reflecting the lack of rock types of intermittent chemical composition
(Fig. 7.11). Following Ashwal (1993), this separation of data points is a common feature of
unmetamorphosed anorthosite massifs. In clear contrast, metamorphosed anorthosite complexes
display an apparent chemical continuity between the anorthosite and granitoid suites, which the
Ashwal (1993) interprets to have been caused by secondary processes.
8 Oxygen isotope constraints
135
The utility of stable isotopes as monitors for the source rocks and processes like crustal
assimilation and hydrothermal alteration of anorthosites and related rocks has been demonstrated
in various studies (Taylor, 1968; Valley & OŃeill, 1982, Morrison & Valley, 1988; OĆonnor
& Morrison, 1999; Peck & Valley, 2000).
The stable isotope analyses were carried out at the Geowissenschaftliches Zentrum,
Abteilung Isotopengeologie, Universität Göttingen. For analyses of the oxygen isotopic
composition of the anorthositic suite, plagioclase was separated from 8 samples of the dark
anorthosite and 6 samples of the white anorthosite suite. In addition, one olivine separate of a
leucotroctolite sample has been analysed. Oxygen isotopic compositions of the felsic suite were
determined for K-feldspar and plagioclase separates for one quartz-syenite and 4 syenite
samples. The obtained plagioclase and K-feldspar 18O values are a good proxy for the bulk rock
δ18O since both, the anorthosites and syenites are almost entirely composed of feldspar. A
description of the analytical techniques is given in the Appendix A.2.6.1. Analytical results are
listed in Table A.5.5.1 in the Appendix A.5.5.
8.1 Anorthositic suite
Almost all plagioclase separates from pyroxene-bearing anorthosite (Ku-97-31, Ku-9792a), olivine-bearing anorthosite and leucotroctolite (Ku-97-95, Ku-97-125, Ku-98-220) and
leucogabbronorite (Ku-98-128, Ku-98-222) of the dark anorthosite suite exhibit a narrow range
in δ18O (5.61-6.13 ‰; Fig. 8.1), well in accordance with typical values of 5.8-7.6 ‰ for
magmatic anorthosite from occurrences world-wide (e.g. Ashwal, 1993). The obtained
18
O/16O
ratios are similar to those from typical basalts and gabbros and are thus consistent with
uncontaminated mantle-derived parental melts (Fig. 8.2). Pyroxene from leucotroctolite Ku-9795 of the dark anorthosite suite is distinctly lower in δ18O (3.07 ‰) when compared to
plagioclase (5.90 ‰) from the same sample, suggesting subsolidus diffusional exchange.
Just one leucotroctolite sample (Ku-97-105), taken from a fault zone near the border
between the KIC and the amphibolite-facies rocks of the Epupa Complex, has a considerably
lower plagioclase δ18O of 3.19 ‰. The low oxygen isotopic value is most probably related to an
exchange with heated meteoric or hydrothermal fluids, restricted to the weak contact zones
between the anorthosite massif and its country rocks. This interpretation is in excellent
136
agreement with the occurrence of hydrothermally formed fibrous actinolite in the sample, which
is a late-stage replacement-product of garnet in the coronas textures surrounding olivine.
Extremely low δ18O values of down to -6‰ have also been analysed for the Boehls Butte, Idaho
and Bitterroot, Montana anorthosite complexes (Mora & Valley, 1988) which the authors
interpret to have resulted from subsolidus alteration of these rocks.
6
5
6
plagioclase, white anorthosite
5
4
4
3
3
2
2
1
1
0
-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
0
plagioclase, dark anorthosite
-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
18
δ OSMOW
δ 18OSMOW
Fig. 8.1: Oxygen isotope composition histograms of magmatic plagioclase from the white and the dark anorthosite
suite of the Kunene Intrusive Complex.
Three of five samples of the white anorthosite (Ku-97-33, Ku-97-33a, Ku-98-60) display
plagioclase δ18O contents of 5.35-6.14 in the range of unaltered plagioclase from the dark
anorthosite (Fig. 8.1). Thus, a pervasive alteration by circulating crustal fluids or by meteoric
water seems rather unlikely. A more convincing model is, that the alteration of the white
anorthosite results from autometasomatism or metasomatism by fluids released by the younger
anorthositic melts of the dark anorthosite suite, since only fluids like these would have no effects
on the isotopic composition of the plagioclases. This interpretation is in excellent agreement with
the preservation of magmatic REE patterns of the white anorthosite.
One conspicuously high δ18O value of 7.30 has been determined for plagioclase from one
white anorthosite sample (Ku-97-33b), which is ~ 1.4 ‰ higher than the values found in the
"normal", magmatic plagioclase of the white and dark anorthosite. This date most probably
reflects contamination with crustal calc-silicate material, since the homogeneous and lower δ18O
values of almost all other samples strongly argue against a regional interaction with crustalderived, high δ18O fluids. In contrast, one brecciated and completely re-crystallized white
anorthosite sample (Ku-98-45), taken from a major fault zone, exhibits a lower δ18O value of
137
2.36, which agrees well with the one obtained for the altered leucotroctolite sample from a
similar tectonic setting.
magmatic plagioclase of the
anorthosite suites
basaltic rocks
low δ18O plagioclase of the
anorthosite suites
meteoric waters
high δ18O plagioclase of the
white anorthosite suite
metamorphic rocks
magmatic feldspar of the
felsic suite
granitic rocks
50
40
30
20
10
18
0
-10
-20
-30
-40
-50
δ OSMOW
Fig. 8.2: Comparison of the δ18O values of feldspar from the anorthositic suites of the KIC and the felsic suite with
oxygen isotope data from important oxygen-containing compounds of the lithosphere after Hoefs (1997).
In conclusion, the intermediate plagioclase δ18O values (average: 5.88 ± 0.19 ‰ δ18O) of
both the dark and the white anorthosite suite are interpreted to be magmatic in origin, lower
plagioclase δ18O values (2.36-3.19 ‰ δ18O) to have resulted from the influx of meteoric waters
restricted to faults and the higher δ18O value (7.30 ‰ δ18O) to have been caused by crustal
contamination.
8.2 Felsic suite
The δ18O values of 4 K-feldspar separates from both the quartz-syenite (Ku-98-40) and the
syenites (Ku-99-13, Ku-99-20, Ku-99-21) falls in a restricted range of 7.22-7.92 ‰ (mean: 7.50
± 0.26 ‰; Fig. 8.3). Almost identical values have been obtained for plagioclase separates of the
respective samples (7.20-7.52 ‰; mean: 7.36 ± 0.16 ‰; Fig. 8.3), suggesting equilibrium
crystallization of the two feldspars. The obtained values are interpreted to be magmatic, since the
small shift of δ18O values and the coincidence of the 18O/16O ratios of K-feldspar and plagioclase
in the syenites are inconsistent with low-temperature alteration. It has to be mentioned, however,
that the average δ18O value of feldspar from the syenites is about 1.6 ‰ higher than the average
138
plagioclase δ18O of the anorthosites. Following Taylor & Sheppard (1986) there is indeed a
general increase in δ18O with increasing SiO2 during closed-system crystal fractionation, since
silica tends to concentrate the heavy oxygen isotope.
6
6
K-feldspar, syenite
5
4
4
3
3
2
2
1
1
0
plagioclase, syenite
5
-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
0
-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
18
18
δ OSMOW
δ OSMOW
Fig. 8.3: Oxygen isotope composition histograms of magmatic K-feldspar and plagioclase from the syenite of the
felsic suite.
However, following these authors, even for an increase of 75% for silica the concomitant
18
O
18
increase is generally < 1 ‰, for a given fractionation temperature of > 900°C. This O increase
9
8
crustal
contamination
18
δ OSMOW
7
6
5
anorthositic suite
4
subsolidus
alteration
3
2
1
40
45
50
felsic suite
Felsic suite
quartz-syenite
syenite
Anorthositic suite
white anorthosite
leucogabbronorite
leucotroctolite (and A,ol)
55
60
65
70
SiO2 (wt.%)
Fig. 8.4: Plot of δ18O against SiO2 for the anorthositic rocks of the KIC and the felsic suite (see text for discussion).
is significantly lower than that obtained for the granitic and anorthositic suites which, at the same
time, only display a relative silica increase of ~ 30 % (Fig. 8.4). The great difference in the
139
magmatic 18O/16O ratios of feldspars from syenites and anorthositic rocks rather suggests a noncomagmatic relationship of the two suites. A possible explanation for the gap between the
isotopic compositions is the incorporation of crustal components into the granitic melts. The
rather homogeneous δ18O values of the syenitic feldspars point to a crustal origin or mixed
mantle and crust sources for the felsic suite.
140
9 Pressure-temperature conditions
The P-T conditions for magmatic crystallization and subsolidus re-equilibration of the
anorthositic suite (Drüppel et al., 2001) and the felsic suite can be estimated from the chemical
compositions and textural relationships of their respective mineral assemblages. The results of
the various calculations are summarized in the P-T plot of Fig. 9.3.
9.1 Thermobarometry of the anorthositic rocks
9.1.1 Magmatic textures
The following magmatic crystallization sequence was established for samples of the dark
anorthosite suite:
cumulus plagioclase + Fe-Ti oxides + olivine →
intercumulus Fe-Ti oxides + olivine + orthopyroxene + clinopyroxene →
late-magmatic amphibole + biotite
Plagioclase and part of the Fe-Ti-oxides and olivine are of cumulus origin, followed by
intercumulus Fe-Ti oxides, olivine, orthopyroxene and clinopyroxene. If both, olivine and
pyroxenes are present, the latter commonly form rims around olivine. Late-magmatic rims of
amphibole and biotite, crystallized from pockets of trapped melt, surround intercumulus
pyroxenes and the Fe-Ti oxides.
Orthopyroxene-clinopyroxene thermometry
In order to constrain the crystallization temperatures of the anorthositic rocks, the temperatures
of 3 pairs of coexisting interstitial ortho- and clinopyroxene in the dark anorthosite were
calculated using the thermometer formulation of Brey & Köhler (1990; their equation 9), based
on the transfer reaction of enstatite in Opx and Cpx. For an assumed pressure range of 6-8 kbar
(for a detailed discussion on the reasons for the pressure assumption see Chapter 9.3),
temperatures of 730-780°C (Ku-98-52) and 740-745°C (Ku-98-221b) were calculated (Table
9.1). The accuracy of the calculation in natural systems is ± 16°C. The estimated temperatures
resemble those derived with the pyroxene-solvus diagram of Lindsley (1983), with the solvus
141
isotherms of coexisting orthopyroxene and clinopyroxene displayed in the quadrilateral pyroxene
system Di-Hd-En-Fs.
The obtained temperature range is lower than expected for magmatic phases and lies even
below the temperatures calculated from amphibole compositions (see below). This points to a
subsolidus re-equilibration, that strongly affected the Ca-Mg-Fe-exchange in the mafic silicates.
rock type
sample
assemblage
B&K
7.0 kbar
7.5 kbar
leucotroctolite
Ku-98-52
opx2-2-cpx1-7
opx1-5-cpx1-1
776°C
727°C
778°C
729°C
leucotroctolite
Ku-98-221b
opx2-1-cpx1-2
742°C
744°C
Table 9.1: Representative T estimates for magmatic orthopyroxene-clinopyroxene pairs of leucotroctolite samples
of the dark anorthosite suite (B&K Brey & Köhler, 1990).
Hornblende thermometry
For the determination of the T conditions during the late-magmatic crystallization of the
anorthositic rocks the amphibole thermometer of Helz (1973, 1976) was used. Experimental
studies of Helz (1973), at conditions of 700-1000°C and 5 kbar, have shown that the Ti content
of hornblende increases with temperature and is independent of the whole-rock composition if
ilmenite is present and the oxygen fugacity is near the QFM buffer. According to Otten (1984)
the thermometer is applicable to amphiboles of both magmatic and subsolidus origin. In the
presence of ilmenite the amphiboles crystallized from the interstitial magma will contain the
maximum amount of Ti, and the thermometer will yield the crystallization temperature. In
contrast, the Ti-content of amphiboles formed during subsolidus reactions will strongly depend
on the local availability of Ti so that the respective amphiboles may not contain the maximum
possible amount of Ti. Conformably, the thermometer will yield minimum temperatures for
subsolidus amphiboles.
In the dark anorthosite suite of the KIC eighteen amphiboles of presumably latemagmatic origin surrounding ilmenite, were investigated. The number of Ti cations per formula
unit (23 O) varies between 0.27 and 0.40, pointing to crystallization temperatures of 951-986°C.
142
The absence of exsolved Ti-phases like ilmenite or rutile demonstrates that the original Ti
content is preserved in the amphiboles. Consequently, the calculated high temperatures are
interpreted to give the late-magmatic crystallization temperature of the dark anorthosite suite.
These temperature estimates are in good agreement with temperatures determined for similar
amphiboles of the Artfjället dolerites, Sweden (Otten, 1984) and the Frankenstein gabbro,
Odenwald, Germany (Kreher, 1992), which are supposed to be near-solidus.
Ilmenite-magnetite thermometry and oxygen barometry
The estimation of temperatures and oxygen fugacities for unexsolved, cogenetic ilmenite and
magnetite was performed using the Ilmss-Mtss thermometer-oxy-barometer of Spencer &
Lindsley (1981), based on hydrothermal exchange-equilibrium experiments in a temperature
range of 500-1000°C by Spencer & Lindsley (1978). The thermometer is applicable for a wide
variety of igneous and metamorphic rocks. The choice of pressure conditions does not influence
the obtained results.
The calculation yields temperatures in the range of 680-730°C with corresponding
oxygen fugacities of 10-17-10-19, lying near the QFM buffer (Table 9.5; Fig. 9.2). Since ilmenite
and magnetite commonly predate amphibole, the temperatures indicate re-equilibration of the
Fe-Ti oxides as in the case of coexisting pyroxenes.
9.1.2 Coronitic mineral assemblages
In some of the leucotroctolite samples, the primary magmatic textures are partially obliterated by
garnet-bearing corona structures around interstitial olivine and Fe-Ti oxides, pointing to a
subsolidus re-equilibration of the rocks. Thus, the obtained quantitative P-T estimates should
shed some light on the post-emplacement evolution of the Kunene Intrusive Complex.
Garnet-orthopyroxene thermometry
Temperatures during corona formation around olivine were estimated with the thermometers of
Lee & Ganguly (1988) and Bhattacharya et al. (1991), based on temperature-dependent Fe-Mg
exchange between coexisting garnet and orthopyroxene, expressed by the reaction:
143
1/3 Mg3Al2Si3O12 + FeSiO3 ↔ 1/3 Fe3Al2Si3O12 + MgSiO3.
pyrope
ferrosilite
almandine
(9-1)
enstatite
The reaction has been calibrated as a thermometer by two independent experimental studies
(Harley, 1984; Lee & Ganguly, 1988). The calibration of Bhattacharya et al. (1991) is based on
the experimental data of Lee & Ganguly (1988), combined with the better constrained values of
mixing parameters for the pyrope-almandine binary of Geiger et al. (1987). The Harley (1984)
calibration has been excluded, since it is very sensitive to retrograde re-equilibration and thus
mostly yields unrealistically low temperatures.
sample
assemblage
L&G
B
7.0 kbar
7.5 kbar
7.0 kbar
7.5 kbar
gt1-opx1
680°C
683°C
663°C
667°C
gt2-opx2
676°C
679°C
658°C
663°C
Ku-97-105
gt1-opx1
632°C
635°C
620°C
624°C
Ku-98-221a
1-gt1-opx1
774°C
777°C
732°C
737°C
2-gt1-opx2
747°C
750°C
710°C
714°C
3-gt1-opx1
759°C
762°C
716°C
719°C
3-gt2-opx2
765°C
768°C
725°C
730°C
3-gt1-opx3
800°C
803°C
757°C
763°C
3-gt2-opx2
766°C
768°C
729°C
734°C
5-gt1-opx2
854°C
857°C
803°C
808°C
Ku-97-104
Ku-98-221b
Table 9.2: Representative T estimates for garnet-orthopyroxene pairs of Type I coronas in leucotroctolite samples of
the dark anorthosite suite (Thermometer calibrations: B Bhattacharya, 1991; L&G Lee & Ganguly, 1988).
For the applied thermometers a pressure range of 7.0 – 7.5 kbar was assumed (for a
detailed discussion on the reasons for the pressure assumption see Chapter 9.3), yielding a range
of temperatures for the four analysed samples, i.e. 630-860°C (Lee & Ganguly, 1988) and 620810°C (Bhattacharya et al., 1991; Table 9.2). However, temperatures obtained for garnetorthopyroxene pairs from individual samples are quite homogeneous, ranging from 632-635°C in
sample Ku-97-105, 676-683°C in sample Ku-97-104, 747-777°C in sample Ku-98-221a and
766-857°C in sample Ku-98-221b. Remarkably, in the same direction the modal amount as well
as the grain sizes of garnet increase, indicating that the garnets from samples Ku-97-104 and –
144
105 may have re-equilibrated, following diffusional exchange with the bordering Fe-Mg silicates
(e.g. orthopyroxene, actinolite), even though continuous garnet formation under decreasing
temperatures during cooling cannot be ruled out, as will be discussed in some detail below.
Garnet-orthopyroxene-plagioclase-quartz barometry
Pressures of the corona formation in the leucotroctolite samples were calculated using the garnetorthopyroxene-plagioclase-quartz barometer, based on the net-transfer reaction Pl + Opx = Grt +
Qtz. Two barometers are applicable for the P calculation of granulite-facies metabasites and
charnockites, based on the following two end-member reactions:
Mg2Si2O6 + CaAl2Si2O8 = 2/3 Mg3Al2Si3O12 + 1/3 Ca3Al2Si3O12 + SiO2
enstatite
anorthite
pyrope
grossular
quartz
Fe2Si2O6 + CaAl2Si2O8 = 2/3 Fe3Al2Si3O12 + 1/3 Ca3Al2Si3O12 + SiO2
ferrosilite
anorthite
almandine
grossular
(9-2)
(9-3)
quartz
For the pressure estimation based on the Mg-exchange reaction (9-2) the barometer calibrations
of Newton & Perkins (1982), Powell & Holland (1988) and Eckert et al. (1991) were used,
whereas pressure estimates following the Fe-exchange reaction (9-3) were calculated with the
barometer calibrations of Bohlen et al. (1983) and Moecher et al. (1988).
For an assumed intermediate temperature of 750°C pressures of 6.3-8.1 kbar, 5.8-7.4 kbar
and 5.8-7.6 kbar were calculated using the calibrations of Newton & Perkins (1982), Powell &
Holland (1988) and Eckert et al. (1991), respectively. In contrast, the pressures obtained for the
Fe-exchange barometer calibrations of Bohlen et al. (1983) and Moecher et al. (1988) for the
same temperature range yield distinctly higher values of 9.2-11.2 kbar and 9.8-11.2 kbar,
respectively (Table 9.3). In addition, the pressure estimates calculated from reaction (9-3) are
very sensitive to changes in the assumed T, reflecting a considerably steeper slope of the Fe end
member reaction when compared to reaction (9-2). Moreover, it has to be mentioned, that the
calculated P values are maximum pressures because quartz is present in only minor proportions
and the applied net-transfer reactions are shifted to lower pressures with a decreasing activity of
SiO2. Thus, the lower P ranges derived by reaction (9-2) are considered as more reasonable and
trustworthy results.
sample
assemblage
700°C
145
N&P (Mg)
750°C
800°C
700°C
H&P (Mg)
750°C
800°C
E (Mg)
750°C
700°C
800°C
Ku-97-104 gt1-opx1-pl1
gt2-opx2-pl2
6.2 kbar
6.4 kbar
6.3 kbar 6.4 kbar
6.5 kbar 6.6 kbar
5.7 kbar
5.8 kbar
5.8 kbar 5.8 kbar
5.9 kbar 5.9 kbar
5.8 kbar 5.8 kbar 5.9 kbar
Ku-97-105 gt1-opx1-pl1
6.4 kbar
6.5 kbar 6.6 kbar
5.8 kbar
5.9 kbar 5.9 kbar
Ku-98-221a 1-gt1-opx1-pl1
2-gt1-opx2-pl1
3-gt1-opx1-pl1
3-gt2-opx2-pl2
7.4 kbar
6.9 kbar
6.9 kbar
7.1 kbar
7.5 kbar
7.0 kbar
7.1 kbar
7.0 kbar
7.7 kbar
7.1 kbar
7.2 kbar
7.4 kbar
6.9 kbar
6.4 kbar
6.4 kbar
6.6 kbar
7.1 kbar
6.4 kbar
6.5 kbar
6.8 kbar
7.0 kbar
6.4 kbar
6.5 kbar
6.7 kbar
Ku-98-221b 3-gt1-opx3-pl1
3-gt2-opx2-pl2
5-gt1-opx2-pl1
7.4 kbar
7.0 kbar
7.9 kbar
7.6 kbar 7.8 kbar
7.3 kbar 7.5 kbar
8.1 kbar 8.3 kbar
6.9 kbar
6.6 kbar
7.2 kbar
7.1 kbar 7.3 kbar
6.7 kbar 6.9 kbar
7.4 kbar 7.5 kbar
sample
assemblage
700°C
B (Fe)
750°C
800°C
7.2 kbar
6.5 kbar
6.6 kbar
6.9 kbar
700°C
M (Fe)
750°C
7.2 kbar
6.5 kbar
6.6 kbar
6.8 kbar
7.4 kbar
6.6 kbar
6.7 kbar
7.0 kbar
800°C
Ku-97-104
gt1-opx1-pl1
gt2-opx2-pl2
Ku-97-105
gt1-opx1-pl1
Ku-98-221a 1-gt1-opx1-pl1
2-gt1-opx2-pl1
3-gt1-opx1-pl1
3-gt2-opx2-pl2
9.6 kbar 10.2 kbar
8.7 kbar 9.2 kbar
8.7 kbar 9.3 kbar
8.9 kbar 9.4 kbar
10.8 kbar
9.8 kbar
9.8 kbar
9.9 kbar
Ku-98-221b 3-gt1-opx3-pl1
3-gt2-opx2-pl2
5-gt1-opx2-pl1
Table 9.3: Representative P estimates, calculated using the garnet-orthopyroxene-plagioclase-quartz barometer for
garnet-orthopyroxene-plagioclase-quartz assemblages of Type I coronas in leucotroctolite samples of the dark
anorthosite suite. (Barometer calibrations: B Bohlen et al., 1983; E Eckert et al., 1991; M Moecher et al., 1988; N&P
Newton & Perkins, 1982, P&H Powell & Holland, 1988).
Like the T estimates for coexisting garnet and orthopyroxene of the leucotroctolite
samples the calculated P for the Tref of 750°C range can be subdivided into two groups, with the
highest pressures of 6.4-8.1 kbar (7.3 ± 0.9 kbar) obtained for samples Ku-98-221a and Ku-98221b, which at the same time yield highest Grt-Opx temperatures (750-860°C), whereas lower
pressures were calculated for samples Ku-97-104 and Ku-97-105 (6.2 ± 0.4 kbar), corresponding
to lower Grt-Opx temperatures (630-680°C; Table 9.3).
146
The most plausible explanation for these results is a garnet formation in at least two
stages, the first generation predating and the second one post-dating a weak exhumation of the
anorthosite body from ~7.3 (stage I) to ~6.2 kbar (stage II). Such an uplift of about 3 km (~1
kbar) would be in excellent accordance with the currently preferred tectonic setting of
Proterozoic massif-type anorthosites, i.e. their emplacement in extensional environments
(Ashwal, 1993). It has to be mentioned, however, that the low P and T estimates of Ku-97-104
and Ku-97-105 may also result from a later re-equilibration of their coronitic mineral
assemblages with respect to Fe, Mg and Ca.
Garnet-biotite thermometry
The thermometer is based on the temperature-dependent Mg-Fe exchange between coexisting
garnet and biotite, expressed by the reaction:
Fe3Al2Si3O12 + KMg3AlSi3O10(OH)2 = 2 Mg3Al2Si3O12 + KFe3AlSi3O10(OH)2
almandine
sample
phlogopite
assemblage
Ku-98-221b
4-bt incl-gt1
4-bt1-gt2
2-bt1-gt3
3-bt2-gt3
sample
assemblage
Ku-98-221b
4-bt2-gt1
4-bt1-gt2
2-bt1-gt3
3-bt2-gt3
pyrope
I&M
7.0 kbar
7.5 kbar
541°C
703°C
704°C
698°C
542°C
705°C
706°C
700°C
7.0 kbar
7.5 kbar
620°C
712°C
713°C
710°C
620°C
713°C
713°C
710°C
B
(9-4)
annite
K&R
7.0 kbar
596°C
747°C
748°C
735°C
F&S
7.0 kbar
566°C
727°C
726°C
724°C
P
7.5 kbar
7.0 kbar
7.5 kbar
598°C
749°C
751°C
738°C
573°C
725°C
726°C
720°C
574°C
726°C
727°C
721°C
7.5 kbar
567°C
729°C
729°C
726°C
H&S
7.0 kbar
625°C
788°C
788°C
785°C
7.5 kbar
627°C
790°C
791°C
787°C
Table 9.4: Representative T estimates for garnet-biotite pairs of Type II coronas in leucotroctolite samples of the
dark anorthosite suite (Thermometer calibrations: B Bhattacharya et al., 1992; F&S Ferry & Spear, 1978; H&S
Hodges & Spear, 1982; I&M Indares & Martignole, 1985, K&R Kleemann & Reinhardt, 1994; P Perchuk &
Lavrentéva, 1983).
147
The reaction was calibrated as a thermometer by many authors. For the calculation of the
pressure conditions of the garnet-biotite corona formation in the leucotroctolite sample Ku-98221b, the thermometer formulations of Ferry & Spear (1978), Hodges & Spear (1982), Perchuk
& Lavrentéva (1983), Indares & Martignole (1985), Bhattacharya et al. (1992) and Kleemann &
Reinhardt (1994) were used.
For an assumed pressure range of 7.0-7.5 kbar (for a detailed discussion on the reasons
for the pressure assumption see Chapter 9.3), the calculated temperatures for core compositions
of inclusion-free garnet and matrix biotite range between 660-790°C, the highest values being
obtained with the Hodges & Spear (1982), and the lowest with the Indares & Martignole (1985)
formulations (Table 9.4). In contrast, the T estimate using the composition of poikilitic garnet
and a biotite inclusion deviates significantly from these results by yielding unrealistically low
temperatures of 540-630°C. This low T range most probably results from diffusional Fe-Mg
exchange between the poikilitic garnet and the included biotite, which, however, has not affected
inclusion free garnet and matrix biotite (see Chapter 6.8).
148
9.1.3 Symplectitic mineral assemblages
Ilmenite-magnetite thermometry and oxygen barometry: Temperatures and oxygen fugacities for
symplectitic ilmenite-magnetite pairs were estimated using the two-oxide thermometer of
Spencer & Lindsley (1981).
The calculated temperatures are in the range of 490-570°C, with corresponding oxygen
fugacities of 10-20-10-26, indicating symplectite formation or re-equilibration at temperatures
distinctly lower than those calculated for the interstitial oxide pairs, which however still plot near
the QFM buffer (Table 9.5; Fig. 9.2).
rock type
sample
textural position
T (°C)
log10fO2
leucotroctolite
Ku-97-105
interstitial
728
723
723
727
-17.60
-17.50
-17.50
-17.60
leucotroctolite
Ku-98-221b interstitial
680
-19.50
leucotroctolite
Ku-98-52
symplectitic
487
489
507
-24.90
-26.60
-24.50
leucogabbronorite Ku-98-79
symplectitic
518
-23.90
leucogabbronorite Ku-98-84
symplectitic
574
558
-19.80
-20.80
Table 9.5: Representative T and Log10fO2 estimates for ilmenite-magnetite pairs from anorthositic rock samples of
the dark anorthosite suite.
9.2 Thermobarometry of the syenitic suite
Feldspar thermometry
The determination of temperatures during K-feldspar crystallization of the syenites, followed the
feldspar thermometer of Fuhrman & Lindsley (1988), which is based on a ternary solution model
that accounts for the effects of K in plagioclase and of Ca in K-feldspar. This thermometer is
thus highly applicable for feldspars formed under high temperatures, where the effects of ternary
149
solution become increasingly important. Reintegrated compositions of perthitic K-feldspar from
the syenite have been plotted in the ternary feldspar diagram. Solvi for temperatures of 6001100°C at an assumed pressure of 6.5 kbar were calculated with the program SOLVCALC (Wen
& Nekvasil, 1994), applying the feldspar activity model of Fuhrman & Lindsley (1988).
An
An
6.5 kbar
6.5 kbar
Ku-99-11
beam size: 20 µm; N=589
Ku-99-12
single-point analyses
reintegrated composition
°C
00
11
°C
00
11
°C
00
10
C
0°
90
°C
700
600°
C
800°C
°C
00
10
C
0°
90
°C
800
600
°C
700°C
Or
Ab
Or
Ab
An
6.5 kbar
Ku-99-13
°C
00
11
°C
00
10
C
0°
90
0°C
80
°C
700
600°
C
Ab
Or
Fig. 9.1: T estimates for reintegrated compositions of perthitic K-feldspar from the syenites, plotted in the ternary
feldspar diagram. Solvi at 6.5 kbar were calculated with the program SOLVCALC (Wen & Nekvasil, 1994)
applying the feldspar activity model of Fuhrman & Lindsley (1988).
Feldspar core compositions of the three syenite samples investigated mostly plot between
the 800°C and the 900°C isotherm or slightly below the 800°C isotherm, pointing to
crystallization temperatures of 790-890°C, which are presumably near the liquidus of the
150
syenites (Fig. 9.1). In clear contrast, the rim compositions of the feldspars analysed generally
yield temperatures below 750°C.
Amphibole barometry
The experimental studies of Hammarstrom & Zen (1983, 1985, 1986) have shown that the Al
content in amphibole generally increases with increasing pressure. Based on this observation, the
authors compiled a barometer formulation, based on the amount of Altot (total amount of Al
calculated anhydrous on a 23-oxygen basis) in amphibole from calc-alkaline plutons and
calibrated for a T range of 500-1100°C and P ranges of 1.5-3 kbar and 7-10 kbar, respectively.
The barometer has been extended by Hollister et al. (1987) by including pressures between 3 and
7 kbar. Since the hornblende composition varies significantly with the rock composition,
samples containing plagioclase, quartz, hornblende, biotite, potassium feldspar, magnetite and
titanite are most suitable for pressure calculations (Hammarstrom & Zen, 1986).
The amount of Altot of hastingsite from the syenite samples varies between 1.91-2.10,
corresponding to crystallization pressures of 6.0-7.1 kbar (Table 9.6). A similar pressure range
has been determined for symplectitic amphibole from sample Ku-99-11 (P = 6.6-6.9 kbar), thus
suggesting that the syenite emplacement was followed by almost isobaric cooling under P
conditions of 6.5 ± 0.6 kbar.
Pmin
Pmax
Paverage
σ
N
interstitial
symplectitic
6.3 kbar
6.6 kbar
7.1 kbar
6.9 kbar
6.5 kbar
6.8 kbar
0.21
0.10
21
5
Ku-99-13
interstitial
6.0 kbar
6.8 kbar
6.4 kbar
0.19
32
Ku-99-14
interstitial
6.0 kbar
7.1 kbar
6.6 kbar
0.26
43
rock type
sample
textural position
syenite
Ku-99-11
Table 9.6: P estimates for magmatic amphibole of the syenite, calculated by using the Al in hornblende barometer
calibration of Hollister et al. (1987).
Taken into account the overall uncertainties in conventional geothermobarometry, these
pressure estimates are in good accordance with those obtained for the second stage of corona
formation in the leucotroctolites (6.2 ± 0.4 kbar).
151
Ilmenite-magnetite thermometry and oxygen barometry
Temperatures calculated with the two-oxide thermometer of Spencer & Lindsley (1981) for
interstitial ilmenite and magnetite from the quartz-syenite range between 600-620°C with
corresponding oxygen fugacities of 10-18 (Table 9.7).
rock type
sample
textural position
T (°C)
log10fO2
quartz-syenite Ku-98-40
interstitial
622
624
599
-18.10
-18.00
-17.90
syenite
symplectitic
540
-18.10
Ku-99-11
Table. 9.7: Representative T and Log10fO2 estimates for ilmenite-magnetite pairs from quartz-syenite and syenite of
the felsic suite.
-10
-14
log10fO2
HM
M
QF
-18
MW
-22
-26
-30
400
Felsic suite
interstitial Ilm-Mag pairs
symplectitic Ilm-Mag pairs
Anorthositic suite
interstitial Ilm-Mag pairs
symplectitic Ilm-Mag pairs
O
NN
500
600
700
800
900
T (°C)
Fig. 9.2: Plot of log10ƒO2 against temperature (°C) calculated for coexisting ilmenite-magnetite pairs occurring in
different textural positions in both the anorthositic and the felsic suite (HM hematite-magnetite buffer, MW
magnetite-wuestite buffer, NNO Ni-NiO buffer, QFM quartz-fayalite-magnetite buffer).
The temperature obtained for ilmenite and magnetite in symplectitic intergrowth with
amphibole and plagioclase in the syenite is lower (540°C), but yields a similar value for the
oxygen fugacity (log10fO2: 10-18). In contrast to those of anorthositic ilmenite-magnetite pairs,
these oxygen fugacities are higher than the QFM buffer (Fig. 9.2), indicating a relatively
oxidised environment, which is consistent with the presence of hematite and the absence of
olivine from the felsic rocks.
152
9.3 P-T evolution of the Kunene Intrusive Complex and the temporally associated
felsic suite
In the dark anorthosite suite, plagioclase was the first mineral to crystallize, accompanied and
followed by cumulus to intercumulus magnetite and olivine. The next minerals appearing in the
crystallization sequence are interstitial orthopyroxene, clinopyroxene and ilmenite. Ilmenite as
well as the pyroxenes are partially surrounded by rims of pargasitic to hastingsitic amphibole and
biotite. A late-magmatic crystallization rather than a subsolidus formation of amphibole is
suggested by the following observations: (1) amphiboles of similar composition occur in
different textural positions, i.e. isolated or as rims locally surrounding pyroxene or ilmenite. (2)
In contrast to pyroxene, amphibole frequently contains inclusions of euhedral apatite grains,
pointing to the late-magmatic crystallization of apatite and pargasitic to hastingsitic amphibole
from P2O5-rich residual liquids. (3) A subsolidus replacement of pyroxene seems to be unlikely,
since no crystallographic relationship between ortho- and clinopyroxene and amphibole has been
observed. (4) In contrast to the pargasitic to hastingsitic amphiboles, actinolite formed by
topotactic subsolidus replacement of orthopyroxene includes platelets of ilmenite exsolved from
the precursor orthopyroxene. (5) The high Ti-contents of the pargasitic to hastingsitic
amphiboles are suggestive of igneous crystallization, as subsolidus re-equilibration would
typically result in exsolution of rutile or ilmenite (Otten, 1984). (6) If the analysed amphiboles
have been crystallized under subsolidus conditions, the obtained temperatures of ~970°C would
reflect minimum temperatures of the amphibole formation, due to the restricted availability of Ti
(Otten, 1984). Subsolidus temperatures of > 970°C, however, would necessitate a high-T ductile
deformation event that was not observed. (7) The absence of hydrous silicates in the ilmenitemagnetite-orthopyroxene symplectites contradicts the presence of a H2O-rich fluid phase in the
subsolidus stage.
Following these arguments, the average T of 970°C obtained from the amphibole
thermometer of Helz (1973) is interpreted to be the late-magmatic crystallization temperature of
the dark anorthosite (Fig. 9.3). This late-magmatic crystallization of amphibole must have
occurred after the emplacement of the anorthositic crystal mushes.
After solidification, the formation of a first generation of garnet-orthopyroxene coronas
around olivine and of garnet-biotite coronas around ilmenite occurred under P-T conditions of
around 760 ± 100°C and 7.3 ± 0.9 kbar. Similar temperature ranges of 720-780°C and 680730°C, respectively, are obtained for magmatic orthopyroxene-clinopyroxene and ilmenitemagnetite pairs. These P-T conditions are indicative of an overall re-equilibration of the plutonic
153
mass at conditions near the amphibolite- to granulite-facies transition (Fig. 9.3), as already
described for various massif-type anorthosite complexes (e.g. Martignole & Schrijver, 1971;
Griffin, 1971; Bingen et al.; 1984, Ashwal et al., 1998). Since there is no evidence for a highgrade metamorphic event post-dating the anorthosite emplacement both on the Namibian part
(Drüppel et al., 2001) and the Angolan part of the KIC (e.g. Ashwal & Twist, 1994; Morais et
al., 1998; Mayer et al., 2000) the garnet-bearing corona structures are suggested to reflect
cooling of the anorthosite body rather than a separate event of prograde metamorphism. This
interpretation is mainly based on the following observations: (1) Deformation textures like shear
zones and foliation are absent in the two anorthosite suites, both on the macroscopic and
microscopic scale. (2) Metamorphic textures as well as corona structures have not been observed
in samples of the white anorthosite suite, which obviously predates the dark anorthosite
emplacement and consequently should have been affected by a prograde metamorphic event as
well. (3) The preservation of hydrothermal alteration in the white anorthosite is unlikely in case
of a pervasive high-grade metamorphic overprint. If, however, the alteration of the white
anorthosite took place during retrograde greenschist-facies metamorphism along major E-W
trending faults, these should extend from the KIC to the surrounding basement. This has never
been observed. (4) Garnet-bearing corona structures have not yet been observed in most of the
KIC and occur only in a few leucotroctolite samples of the dark anorthosite suite. Even within
single samples they are found just sporadically. (5) On the microscopic scale, igneous textures
are preserved in all investigated samples, whereas metamorphic effects like granoblastic textures,
expulsion of Fe-Ti oxide inclusions from plagioclase or reduction of the grain size have not been
observed.
Following these arguments, the pressures of around 7.3 ± 0.9 kbar, calculated with the
garnet-orthopyroxene-plagioclase-quartz barometer for the subsolidus formation of garnet and
orthopyroxene, are interpreted to reflect the emplacement depth (c. 22 km) of the anorthositic
crystal mushes. Consequently, pressures of c. 7.3 kbar can also be assumed for the late-magmatic
crystallization of pargasite. In two leucotroctolite samples, bearing minor and very fine-grained
garnet,
garnet
orthopyroxene
thermometry
and
garnet-orthopyroxene-plagioclase-quartz
barometry yielded slightly lower temperature and pressure ranges, suggesting their formation
under PT conditions of 630-680°C and 6.2 ± 0.4 kbar. The PT estimates indicate an exhumation
of the anorthosite massif of about 3 km (~1 kbar) during the end-stages of the anorthosite
emplacement (Fig. 9.3), which is in excellent agreement with the widely held belief that massiftype anorthosites are emplaced in extensional settings. It cannot be ruled out, however, that the
small grain size of garnet in these two leucotroctolite samples favoured retrograde cation
exchange with the bordering orthopyroxene and plagioclase, thus yielding re-equilibration
154
temperatures and pressures. As no symplectitic breakdown of coronitic garnet was observed, it
can be concluded that pressure conditions of about 6-7 kbar outlasted the garnet formation for a
certain period of time. These results invoke an emplacement in the middle crust followed by
near-isobaric cooling of the anorthosite complex.
9.5
9.0
P (kbar)
10.0
H
coronitic (stage I)
H
LG
B
NP
BK
8.5
IM B
BK
LG
SL
8.0
SL
SL
7.5
NP
7.0
syenite
late-magmatic
NP
Ho
6.0
B LG
5.5
5.0
G
rti
NP
6.5
n
Ho
FS SL
B
LG
amphibolite-facies rocks
of the Orue unit
(Brandt et al., 2003)
coronitic (stage II)
symplectitic
B: Bhattacharya et al. (1991), Grt-Opx
BK: Brey & Köhler (1990), Opx-Cpx
4.5
FS: Ferry & Spear (1978), Grt-Bt
H: Helz (1973, 1976), Ti in Hbl
4.0
Ho: Hollister et al. (1987), Al in Hbl
IM: Indares & Martignole (1985), Grt-Bt
3.5
LG: Lee & Ganguly (1988), Grt-Opx
NP: Newton & Perkins (1982), Grt-Opx-Pl-Qtz
3.0
SL: Spencer & Lindsley (1981), Ilm-Mag
Grt-in: Green & Ringwood (1967)
2.5
2.0
450
T (°C)
500
550
600
650
700
750
800
850
900
950
1000 1050
Fig. 9.3: P-T evolution of the dark anorthosite suite of the Kunene Intrusive Complex (see text for detail). The
results of conventional geothermometry and geobarometry are indicated as the respective maximum and minimum
estimates. For comparison, the P-T conditions of the bordering amphibolite-facies Orue Unit (Brandt et al., 2003)
and the P conditions of the emplacement and cooling of the syenite suite are marked. (arrow suggested subsolidus
P-T path for the dark anorthosite suite, grey stippled line experimentally determined reaction curve for plagioclaseolivine reaction and garnet-formation in olivine-tholeiites by Green & Ringwood, 1967).
155
The corona formation is followed by the development of ilmenite-magnetiteorthopyroxene symplectites at the expense of olivine and igneous Fe-Ti oxide following reaction
(6): The formation of these symplectites necessitates that the QFM buffer curve was reached or
crossed with decreasing temperature. In the case of the dark anorthosite suite, the calculated
f(O2)-T path starts from 10-17-10-19, 680-730°C of the orthomagmatic ilmenite-magnetite pairs
down to 10-20-10-26, 490-570°C calculated for the symplectites. On isobaric cooling, the QFM
buffer curve is crossed, if a CO-CO2-bearing fluid phase is involved, since the ƒ(O2)-T curves of
the solid buffer assemblages are transsected by solid-gas buffer curves like graphite-CO-CO2 (cf.
Fig. L-20 in Lindsley, 1976). Support for this model is given by the presence of graphite
precipitates on small fractures observed in one dark anorthosite sample. Both the orientation
relationship between symplectitic orthopyroxene, magnetite and ilmenite, and the respective
chemical compositions are compatible with those described by Barton & van Gaans (1988) and
Barton et al. (1991), suggesting a subsolidus formation during cooling under dry conditions. The
final stage of the late-magmatic evolution is marked by the formation of actinolite in the garnetbearing coronas around olivine and the alteration of olivine to talc, magnetite and carbonate.
In the (quartz-)syenites of the felsic suite potassium feldspar, plagioclase and part of the
hastingsite/clinopyroxene crystallized first, followed by interstitial hastingsite, Fe-Ti oxides,
which may in turn be surrounded by narrow rims of titanite, which may additionally occur as
discrete, anhedral grains together with a second generation of potassium feldspar (microcline).
During the subsolidus re-equilibration of the syenites interstitial hastingsite was partially
replaced by hastingsite-ilmenite-magnetite-plagioclase symplectites. A weak hydrothermal
alteration of the syenites caused the replacement of clinopyroxene by fibrous ferro-hornblende to
ferro-actinolite and/or epidote-calcite, the marginal alteration of hastingsite to chlorite, the
sericitization of the feldspars and the martitisation of magnetite, whereas the recrystallisation of
feldspars to granular plagioclase mosaics testifies to a post-emplacement deformation of the
rocks.
For the magmatic to late-magmatic crystallization of hastingsite pressures of 6.5 ± 0.6
kbar have been obtained, which is similar to those derived for stage II garnet-orthopyroxene
coronas around olivine (6.2 ± 0.4 kbar). These pressure estimates are interpreted to reflect the
emplacement depth of the syenites (~ 19 km; Fig. 9.3), thus suggesting that the intrusion of the
syenites in the middle crust post-dated the ~ 3 km exhumation of the dark anorthosite suite.
Conformably, field observations and age determinations generally confirm an intrusive
relationship between the two suites, with the felsic rocks being younger than the spatially
associated anorthositic rocks (see Part I).
156
Magmatic crystallization of potassium feldspar in the syenites occurred under
temperatures of 790-890°C, which are presumably near the liquidus of the rocks, whereas the
(partially re-crystallized) feldspar rims formed at distinctly lower T of <750°C. Magmatic
ilmenite-magnetite pairs of the quartz-syenite and syenite yield oxygen fugacities of 10-18 and
comparably low temperatures of 600-620°C and 540°C, respectively, which are suggested to
reflect re-equilibration conditions.
For the subsolidus formation of hastingsite-ilmenite-magnetite-plagioclase symplectites
pressures of 6.6-6.9 kbar have been determined, implying that the syenite emplacement was
followed by almost isobaric cooling, that has also been proposed for the dark anorthosite suite.
In summary, the results indicate that the emplacement of the dark anorthosite suite of the
Kunene Intrusive Complex in the middle crust (P = 7.3 ± 0.9 kbar, T = 950-990°C) was followed
by near-isobaric cooling to temperatures <600°C and pressures of 6.2 ± 0.4 kbar. The intrusion
of the dark anorthosite suite was followed by the emplacement of numerous syenite dykes under
PT conditions of 6.5 ± 0.6 kbar and 790-890°C. Like the anorthosites, the syenites cooled under
almost isobaric conditions to temperatures <600°C. A prograde metamorphic event is precluded
for the analysed samples, for reasons discussed in some detail above. This model of mid-crustal
emplacement and cooling is constrained by PT conditions obtained for the bordering
amphibolite-facies Orue Unit by Brandt et al. (2003). For these rocks, which locally even
underwent contact-metamorphism by the anorthositic melts, the authors calculated temperatures
of ~650-750°C at pressures of ~6-8 kbar, which agree well with the PT estimates for the
emplacement and re-equilibration of both the anorthositic and the felsic suite (Fig. 9.3).
10 Discussion
157
10 Discussion
The Kunene Intrusive Complex displays many of the typical features of massif-type anorthosite
complexes as summarized by Ashwal (1993):
•
The KIC has an extraordinary large areal extent of at least 18,000 km2, being composed
of several plutonic bodies.
•
The massif mainly consists of anorthosite, leucogabbronorite and leucotroctolite, whereas
mafic and ultramafic rocks are rare.
•
Minor concentrations of ilmenite-magnetite ore occur in the marginal parts of the
complex.
•
The KIC is associated with minor volumes of K-rich felsic rocks, ranging in composition
from syenite to granite.
•
Plagioclase has an average composition of An50±20 and is generally coarse-grained.
Features like the mortar texture and the block-structure as well as the occurrence of anorthositic
dykes crosscutting the main massif and the lack of layering additionally confirm the massif-type
character of the KIC; a layered intrusion affinity of the KIC can be ruled out.
10.1 Constraints on the source and evolution of the anorthositic and syenitic melts
Much has been speculated about the magmas parental to anorthositic and associated granitic
melts, however, no general consensus has been reached yet (e.g. Ashwal, 1993, for a review). A
first decision to be made is whether the felsic rocks represent late-stage fractionation products
from the magma parental to the anorthosites, or if the two suites are chemically unrelated.
In the case of the Kunene Intrusive Complex, evidence against consanguinity of the two
suites is striking: (1) Field observations show the syenite dykes to be intrusive into the
anorthositic rocks. In addition, granitoid plutons may form isolated bodies, crosscutting the
bordering amphibolite-facies Epembe Unit of the Epupa Complex. (2) The mineral compositions
of the granitoid suite appear to extend the compositional range of the respective minerals from
the anorthositic rocks towards higher Fe/Mg ratios and lower Al of the Fe-Mg silicates and lower
An contents of plagioclase. However, intermediate mineral compositions are missing. Moreover,
the relative changes of the mineral chemistry, displayed by the subsequently emplaced
158
10 Discussion
anorthosite bodies themselves, rather imply an opposite evolutionary trend (towards higher An of
plagioclase and lower Fe/Mg ratios of the Fe-Mg silicates). (3) A compositional gap is developed
between the major and trace element data of the two suites; chemical continuity, as commonly
observed for most metamorphosed anorthosite massifs, is lacking. (4) The felsic rocks are typical
of Proterozoic crustally derived A-type granites (K-rich, ferroan, subalkaline to alkaline,
metaluminous to marginally peraluminous A-type granitoids). (5) Trace element data of the
felsic suite, plotted in spider diagrams, provides evidence for a crustal or at least a mixed crustalmantle source. (6) The syenites do not exhibit ubiquitous negative Eu-anomalies in their REE
patterns, which would, however, be expected from fractionation products of melts that
previously formed extensive plagioclase cumulates. (7) The magmatic feldspar δ18O values from
both the quartz-syenite and the syenites fall in a restricted range of 7.20-7.92 ‰, which,
however, are about 1.6 ‰ higher than the average magmatic plagioclase δ18O of the anorthosites.
Following these arguments, the felsic suite was at least not formed by closed-system
fractionation from the melts parental to the anorthositic rocks of the Kunene Intrusive Complex.
The question then becomes, were the felsic rocks formed (1) by fractional crystallization from
residual liquids of the anorthosites, contaminated by crustal material during their uprise or (2) by
crustal fusion with the heat necessary being provided by the upwelling and emplacement of
mantle-derived melts at the base of the crust. This can not be answered easily, since in either
case the originally dry magmas will be affected by fractional crystallization and crustal
contamination as they ascend and thus both models could account for most of the features
observed for the granitoids. However, judging from the data set so far available, a crustal origin
of the felsic suite is the preferred working hypothesis, since data constraining a mixed mantlecrust source is equivocal, whereas evidence for a crustal source of the investigated syenites and
granites is impressive. However, isotopic work on both rock suites is in progress (Littmann et al.,
in prep.) and will hopefully allow a final decision for one particular case.
There is no consensus on the magmas parental to massif-type anorthosites, but many
investigators suggest that they are broadly basaltic in nature (e.g. Morse, 1968, 1982; Duchesne,
1984; Emslie, 1985; Wiebe, 1986; 1994, Olson & Morse, 1990; Ashwal, 1993; Longhi et al.,
1999). In case of the KIC, the general geochemical characteristics of all anorthositic lithologies
(i.e. high Al2O3, CaO, Sr and Eu, intermediate to high Mg numbers of 0.37-0.74, Ne- and/or Olnormative compositions) are in excellent agreement with a derivation from fractionated basaltic
liquids. Mineral compositions of both anorthosite suites constrain this model, by being generally
in the range of those commonly found in evolved gabbros and basalts. Following this, the
10 Discussion
159
parental melts of the investigated anorthosites are suggested to be fractionated, Sr-enriched,
high-Al gabbroic magmas.
Regarding the petrogenesis of the melts parental to massif type anorthosites, two different
models have attained some popularity: (1) Most investigators suggest a petrogenetic scheme
involving (i) the ponding of mantle-derived, broadly basaltic melts at the mantle-crust boundary,
followed by (ii) extensive crystallization of olivine and highly aluminous pyroxene, with the
residual liquids becoming more aluminous and Fe-rich, until (iii) plagioclase reaches the
liquidus, leading to the formation of plagioclase suspensions in Fe-rich, high-Al gabbroic liquids,
which (iiii) subsequently intrude middle to upper crustal levels (see Ashwal, 1993, for a review).
This model is mainly constrained by the presence of high-Al orthopyroxene megacrysts within
most anorthosite massifs, which testify to a high pressure stage of crystallization (e.g. Emslie,
1975a; Ashwal, 1993). (2) In contrast, the alternative petrogenetic model of Longhi et al. (1999)
suggests the melts parental to massif-type anorthosites to be generated by partial melting of
mafic source regions of the lower crust. With the help of equilibria experiments, Longhi et al.
(1999) demonstrated that mantle-derived basaltic liquids, fractionating high-Al orthopyroxene in
a first step, would reach plagioclase saturation at low normative Si contents and would thus
become Ne-normative with further fractionation and assimilation of lower crustal material. The
authors argue, that without extensive low-P assimilation of at least 20 % of a granitic component,
such parental magmas could not produce the lithological variety observed in many anorthosite
massifs (e.g. Ne-normative troctolitic and gabbroic lithologies to Qtz-normative noritic
lithologies). Since the primitive isotopic compositions of these anorthosite massifs do not permit
such an extensive assimilation of granitic material, Longhi et al. (1999) suggest mafic regions of
the lower continental crust (mafic granulites or foundered mafic plutons) to be more suitable
sources for the liquids parental to massif-type anorthosite complexes. However, no general
agreement has yet been reached on either of these models.
The δ18O values obtained for magmatic plagioclase from both anorthosite suites of the
KIC prove their derivation from mantle-derived magmas, in good accordance with model (1).
Moreover, the restricted chemical composition of the anorthositic lithologies of the Kunene
Intrusive Complex and the general lack of Qtz-normative noritic assemblages does not require
model (2), the melts parental to the anorthosites of the KIC are thus suggested to have formed by
fractionation of a mantle-derived magma ponded near the base of the crust. However, magma
generation by re-melting of mantle-derived gabbroic plutons, which were previously emplaced in
the lower continental crust, can not yet be ruled out. The fact that UHT metamorphism recorded
by granulites of the Epupa Complex is interpreted in favour of magmatic underplating at ~1,500
Ma (Brandt et al., 2003), might provide evidence for model (2) or a combination of both models.
160
10 Discussion
The mineralogical and isotopic data suggests that all studied anorthosite samples
underwent a certain degree of crustal contamination, which, however, was most prominent
during the early intrusion stages of the white anorthosite and leucogabbronorite: (1) Xenoliths of
andradite-rich garnet, incorporated from calc-silicate crustal lithologies, were observed in one
white anorthosite and one leucogabbronorite sample but never in the younger leucotroctolites.
(2) In contrast, late-magmatic biotite and amphibole are abundant phases in almost all
investigated samples, suggesting a crustal contribution to the anorthositic melts of both suites,
that was most possibly achieved during their ascent into middle crustal levels. (3) Highest values
of Fetot/(Fetot+Mg), Na, K, P and Ba are recorded by the white anorthosite and the
leucogabbronorite and may indicate a granitic component. (4) Regarding their REE patterns
white anorthosite samples display a slight enrichment in the light REE and a relative decrease in
the heavy REE when compared with the corresponding lithologies of the dark anorthosite suite,
supporting evidence for an elevated amount of crustal contamination of these rocks. (5) Last but
not least, oxygen isotope data for one white anorthosite sample provides clear evidence for
incorporation of crustal material.
To summarize, the crustal-derived felsic and the mantle-derived anorthositic suite are
suggested to be coeval but not consanguineous. Their spatial and temporal association can be
accounted for, if the heat necessary for crustal melting is provided by the anorthositic melts. The
studied samples of the two anorthosite suites are representative of distinct intrusions, which
underwent different degrees of crustal contamination, that, however, was most extensive during
the early intrusion stages of the white anorthosite and the leucogabbronorite.
10.2 Petrogenetic model
Mid-Proterozoic continental within-plate magmatism of anorthositic and associated felsic melts
is recognized as an essential part of the early continental evolution of the Earth (e.g. Anderson,
1983; Ashwal, 1993). These rocks are widespread in what were probably large continental
landmasses. The tectonic environment is generally considered to be extensional, even though
their precise tectonic setting is yet under discussion (see Chapter 3.5.1.1). The most favoured
model is a continental rift analogy, which may account for both, the bimodality of anorthositegranitoid suites and the apparent linearity of anorthosite massifs.
10 Discussion
161
The KIC is a broadly N-S elongated body, straddling the Namibian-Angolan border. In
the Namibian part of the KIC, at least two distinct intrusive bodies were distinguished: (1) the
older one mainly consists of light-coloured, strongly altered and tectonically overprinted massive
anorthosites, leucotroctolites and leucogabbronorites, the white anorthosite, and is intruded by
(2) less altered, dark-coloured anorthosites, leucotroctolites and leucogabbronorites, the dark
anorthosite, along E-W-trending extensional faults. Based on the petrographic and geochemical
investigations provided in this study combined with models published previously (e.g. Morse,
1968; Emslie, 1978, 1980; Ashwal, 1993), the following petrogenetic model is suggested for the
evolution of both the KIC and the felsic suite:
1st stage
A significant thermal anomaly leads to the heating and partial melting of the upper mantle and/or
lower crust. The magma ponds at the mantle-crust transition (magmatic underplating), building a
large-scaled magma chamber (Fig. 10.1 a).
2nd stage
Continuing influx of magma and/or assimilation of lower crustal residues causes an expansion of
the magma chamber. The melt increases in Al and Fe/Mg until plagioclase reaches the liquidus,
precipitates and floats, forming plagioclase flotation cumulates. Density contrasts between the
plagioclase-rich mushes and the surrounding country rocks of the lower crust (with density
differences of 0.2-0.6 g/cm3 being estimated by Longhi & Ashwal (1985)) lead to the diapiric
ascent of the crystal mushes and their emplacement in middle crustal levels (Fig. 10.1 a). During
this ascent calc-silicates of the EC are assimilated, as is suggested by andraditic garnet
xenocrysts within white anorthosites.
Late-stage oxide-melt segregations are foundered in the crystal mushes and crystallize the
Fe-Ti ores. Fluid activity, induced by devolatization of younger anorthositic crystal mushes,
leads to a pervasive hydrothermal alteration of the white anorthosite suite, but, however, only
weakly affects its REE.
3rd stage
Deep-crustal, ENE trending lineaments are intruded by a second generation of anorthositic
crystal mush, the dark anorthosite (Fig. 10.1 b). These faults were most probably installed long
before the anorthosite intrusion, since their attitude resembles that of both the foliation of and the
ductile shear-zones within the basement rocks of the Epupa Complex, which were formed during
the Eburnian orogeny at ~ 1.8-1.6 Ga (Brandt et al., 2003). Remarkably, however, the second
162
10 Discussion
Fig. 10.1: Schematic petrogenetic model for the Kunene Intrusive Complex and the associated felsic suite (see text
for details).
10 Discussion
163
generation of anorthositic crystal mushes never intruded into the bordering basement rocks but
solely occurs within the white anorthosite massif, concentrated in its central parts. It can thus be
concluded that the white anorthosite massif itself behaved as a major zone of weakness and was
not fully crystallized during the intrusion of the dark anorthosite.
This model also accounts for the decrease in crustal contamination with continuous
intrusion, since younger melts most probably used similar pathways like the previously
ascending crystal mushes and hence incorporated older anorthositic rather than crustal material.
The emplacement of the dark anorthosite in the middle crust (~ 22 km) was followed by a period
of near-isobaric cooling, leading to a re-equilibration of the magmatic mineral assemblages
under conditions near the amphibolite- to granulite-facies transition.
4th stage
Continued heating of the lower crust finally causes crustal anatexis, hereby forming silicic melts
parental to the granite and syenite of the felsic suite (Fig. 10.1 b). Like the dark anorthosites, the
felsic melts predominantly use E-W trending weakness zones for their ascent. The fact, that the
granitoids are concentrated near the boundaries between the Epupa Complex and the anorthosite
massif, suggests that at least part of the felsic melts are formed by contact-metamorphic anatexis
of the neighbouring basement rocks during the ascent of the anorthositic crystal mushes. The
dyke-like emplacement of the syenites and granites in the middle crust (~ 19 km) is followed by
isobaric cooling.
An extensional setting like the one described could also account for the uplift of the deepcrustal granulites of the Epembe Unit into mid-crustal levels along E-W trending sub-vertical
faults (Brandt et al., 2003). In addition, the emplacement and crystallization of the large volumes
of anorthositic and felsic melts during the period of ~ 1.38-1.35 Ga must have caused a
significant heating of the lower and middle crust (regional "contact-metamorphism"), and may
thus account for the comparably young ages of 1.37-1.33 Ga constrained for the neighbouring
amphibolite-facies rocks of the Epupa-Complex by Seth et al. (in prep.).
164
11 Comparison with other parts of the KIC
11 Comparison with other parts of the KIC
As it is evident from Fig. 1.2 the KIC may be subdivided in at least two major structural parts.
This interpretation is in accordance with the satellite imagery interpretation of Ashwal & Twist
(1994) who recognized at least twelve distinct anorthosite bodies with probably individual
petrogenetic evolution and consequently interpreted the KIC as a composite massif-type
anorthosite. It has to be mentioned however that, due to difficult access during the civil war,
most studies of the northern part of the KIC are based on old field data and sampling (Carvalho
& Alves, 1990, 1993; Silva, 1990, 1992). Recent detailed field work in combination with
petrographical investigations is lacking for most of the Angolan part of the KIC, except restricted
areas around Lufinda, Dongue and Chiange in Angola (Ashwal & Twist, 1994; Morais et al.,
1998, Slejko et al., 2002).
In the Angolan part granitoids are present in minor proportions, mainly represented by
irregular bodies of rapakivi-type granite, graphic granite, leucocratic microgranite and sparse
mangerite dykes (Morais et al., 1998). Like the felsic rocks investigated in this study, these
granitoids are interpreted to be coeval but not consanguineous with the Kunene Intrusive
Complex (Morais et al., 1998). In case of the mangerite dykes this model has been constrained
by age determinations of both the felsic rocks and the KIC (Mayer et al., 2000).
The magmatic features of the leucotroctolite and anorthosite samples of the dark
anorthosite suite investigated in this study resemble those reported for the northern part of the
KIC (Simpson, 1970; Vermaak, 1981; Silva, 1990, 1992; Ashwal & Twist, 1994; Morais et al.,
1998). However, subsolidus reaction textures like garnet-bearing corona structures of the dark
anorthosite have not been described until now. In the Lufinda-Dongue area, the presence of
cogenetic dolerites and mangerites is interpreted in favour of a shallow emplacement level of the
anorthosite (Morais et al., 1998; Slejko et al., 2002), with pressures of 3-5 kbar being estimated
by Slejko et al. (2002) for the subsolidus re-equilibration of pyroxenes. In clear contrast, the
present study of the dark anorthosite of the Zebra Mountains is indicative of emplacement in the
middle crust followed by isobaric cooling. It has to be mentioned, however, that the two
respective study areas are at least 150 km apart from each other (Fig. 1.2). Therefore, detailed
field work, petrographic and geochronological investigation of both anorthosite and country
rocks for the intermittent southern Angolan part of the KIC is needed to decide whether the
apparent differences in erosion level are due to a continuous gradient in exposed crustal depth or
due to offsets by deep fracture zones as for example the Serpa-Pinto lineament or the Kunene
River fault.
12 Introduction
165
PART III: Fenitizing processes induced by ferrocarbonatite
magmatism
12 Introduction
12.1 Carbonatites and fenitization
The majority of the carbonatite complexes known from world-wide localities are predominantly
composed of calcium- and magnesium-rich carbonatites. In these carbonatite centres late-stage
ferrocarbonatite dykes may occur, which are characterized by strong enrichments of Fe, Mg, Ba,
Sr, REE, Th and U. These ferrocarbonatites are commonly interpreted as the end-products of
fractional crystallization and related differentiation processes of the main Ca-Mg carbonatite
body (Le Bas, 1977, 1981, 1989, 1999; Gittins, 1989; Woolley & Kempe, 1989). In clear
contrast, magmatic natrocarbonatite is known only as the extrusive product of the Oldoinyo
Lengai volcano, Tanzania, which, at the same time, is the only active volcano (see Bell & Keller,
1995, for a review). However, the close spatial association of Na- and/or K-rich fenites and
carbonatites (Morogan, 1994) suggests that carbonatites, which are now Ca-Mg-Fe carbonatites,
were alkali-rich when they intruded and expelled alkali-metals and volatiles during their
differentiation and crystallization. This model has been confirmed by experimental studies
(Watkinson & Wyllie, 1971; Freestone & Hamilton, 1980).
Where carbonatitic centres are surrounded by metasomatic aureoles, the composition of
the solidified carbonatite will differ significantly from that of the pristine carbonatite magma due
to material losses and gains during differentiation, crystallization or post-magmatic reequilibration. In this case, the carbonatites only preserve an incomplete memory of their
magmatic history. The investigation of metasomatic mineral reactions at the magma to wall-rock
interface and of fluid inclusions of the fenitizing agents can thus give important information
about the geochemical nature and evolution of the liquids responsible.
The carbonatite dykes of Swartbooisdrif, NW Namibia are an ideal study object since
they are well-exposed, mostly unweathered and suffered only minor subsolidus alteration. These
ferrocarbonatite dykes are not widely known in the international carbonatite literature, even
though their emplacement led to the formation of deep blue sodalite-rich fenites, which are used
as a deposit for ornamental stone, the “Namibia Blue”, since 1960. They differ from all other
known carbonatite occurrences in (1) their exotic fenite mineral assemblage, (2) the uncommon
166
12 Introduction
anorthositic mineralogy of their wall-rocks involved in the metasomatic processes and (3) the
general lack of associated Ca-Mg carbonatites and felsic intrusives. The aim of the present study
is to characterize the magmatic evolution of the carbonatite intrusions and to elucidate the
interrelated metasomatic processes, which led to the formation of the sodalite occurrences.
12.2 Sodalite - a review
Sodalite is one member of the sodalite group, a group of rock-forming framework silicates,
including the end-members sodalite (Na8[Al6Si6O24]Cl2), nosean (Na8[Al6Si6O24]SO4) and
haüyne ((Na,Ca)4-8[Al6Si6O24](SO4,S)1-2). Sodalite differs from the other minerals of the group in
being the most Na-rich member and in containing chlorine as an essential constituent. The name
sodalite (Thomson, 1811) refers to its sodium-rich composition. Variations in the Na content of
sodalite are commonly minor, since the substitution of Na by both K and Ca reaches only a
limited extent. The same holds true for the replacement of Al by Fe3+. In the sodalite variety
hackmanite, sulphur may be present in minor amounts.
Cl
Al
Na
Na
Cl
Al
Al
Cl
Cl
Al
Al
Na
Na
Al
Al
a = 8.87 A
Al
Cl
Fig. 12.1: The framework of tetrahedral groups in the structure of sodalite after Bragg & Claringbull (1965).
The structure of sodalite has been described by Bragg & Claringbull (1965) as an
alumosilicate framework, formed by almost equal amounts of linked SiO4- and AlO4-tetrahedra,
with Al and Si being completely ordered (Fig. 12.1). Six rings of four tetrahedra ||{100} and
eight rings of tetrahedra ||{111} form cage-like cubo-octahedral units. Six-membered rings
define sets of channels, which are occupied by chlorine at their intersections. The Cl- ions
themselves are centred in Na+ tetrahedra. The overall structure is cubic with a = 8.1Å and a
bimolecular unit cell.
12 Introduction
167
The natural occurrence of sodalite is largely restricted to magmatic environments. The
feldspathoid most commonly occurs in SiO2-undersaturated alkaline igneous rocks (nepheline
syenites, phonolites and related rocks), where sodalite is either a magmatic crystallization
product or formed by the hydrothermal alteration of nepheline (Deer et al., 1992). In addition,
sodalite has been described from meteorites (e.g. Berg et al., 1997) and from ejected volcanic
blocks, where it was observed in cavities (e.g. Deer et al., 1992). Sodalite is commonly
associated with nepheline, K-feldspar, albite, cancrinite, natrolite, melanite, calcite and fluorite.
Prominent natural localities of sodalite include phonolites of the Cantal Massif, France (e.g.
Brousse et al., 1969), sodalite foyaite, lujavrite and nauyaite of the Ilimaussaq intrusion, southern
Greenland (e.g. Hamilton, 1964; Sörensen, 1970; Markl et al., 2001), nepheline syenite
pegmatite of the Langesundsfjord, Norway (e.g. Taylor et al., 1967), contact-metamorphic
carbonate and shale sequences of the Oslo Rift, Norway (e.g. Jamtveit et al., 1997), nepheline
syenites, urtites and foyaites of the Lovozero Alkali Massif, Russia (e.g. Sörensen, 1970),
lamprophyres of Kola Peninsula, Russia (e.g. Beard et al., 1996), and sodalite syenite and
nepheline syenite of Mont Saint-Hilaire, Quebec (e.g. Currie et al., 1986).
(1) Experimental investigations of the stability of sodalite have been performed by
several authors (e.g. Wellman, 1970; Barker, 1976; Binsted, 1981; Sharp et al., 1989;
Kostel`nikov & Zhornyak, 1995). Results of the most recent studies of Sharp et al. (1989) and
Kostel`nikov & Zhornyak (1995) are summarized below:
Sharp et al. (1989) studied the stability relations of phases in the system NaAlSiO4-SiO2NaCl at 1 bar and elevated pressures of 7.4-10 kbar over temperature intervals of 1644-1652 K
and 923-1173 K, using natural samples of sodalite-syenites from Mont St-Hilaire, Quebec.
Phases in the system NaAlSiO4-NaCl include halite (NaCl(s)), liquid NaCl (NaCl(l)) and vapour
NaCl (NaCl(v)), nepheline (α-nepheline (289-467K), β-nepheline (467-1180K) and γ-nepheline
(1180-1525K)), carnegieite (δ-carnegieite (298-980K) and β-carnegieite (980-1700K)) and
sodalite. In order to calculate the stability field of sodalite, the authors combined available
thermodynamic data for the above mentioned phases (see Sharp et al., 1989, for details) with
experimental equilibrium reversals for the reaction
6 NaAlSiO4 + 2 NaCl = Na8Al6Si6O24Cl2
nepheline
halite
sodalite
(12-1)
168
12 Introduction
The authors found sodalite to be a high-temperature, low-pressure phase, which is stable even
above the solidus in sodic, silica-undersaturated and NaCl-rich magmas, with the presence of
sodalite constraining the NaCl-activities in the magmas. The experimental results of Sharp et al.
(1989) are illustrated in two selected phase diagrams (Fig. 12.2a, b).
14
14
NaCl(s) NaCl(l)
10
8
Ne(β)
6
Sdl
NaCl(s)
4
Ne(α)
2
0
Ne(γ) Cg
Ne(β)
1000
1200
10
Sdl Cg
NaCl(l)
a
1400
1600
NaCl(s) NaCl(l)
12
Ne(γ)
Sdl NaCl(l)
Pressure (kbar)
Pressure (kbar)
12
Ne(β)
8
Ne(β) + NaCl(l)
= Sdl
NaCl(s)
Sdl
Ne(γ)
Ne(β) Ne(γ)
NaCl(l)
Sdl
6
4
Ne(γ) Cg
Cg
Sdl NaCl(l)
2
1800
0
b
1000
Temperature (K)
1200
1400
1600
1800
Temperature (K)
Fig. 12.2: Stability relation of phases in the system NaAlSiO4-SiO2-NaCl after Sharp et al. (1989). (Cg carnegieite,
Ne nepheline, Sdl sodalite, see Sharp et al., 1989, for details). a) Uses the measured entropy of sodalite at 298 K.
Sodalite-bearing reactions constrained by the 1 bar, 1649-1652 K reversals, except reaction Ne(β) + NaCl(s) = Sdl,
which is determined for high-pressure, 923K and 973 K reversals. b) Includes an excess entropy term for sodalite of
61.7 J/mol٠K. Generated curves coincide with experiments above 973 K.
When expanding the system NaAlSiO4-NaCl to include SiO2, the phases jadeite, albite and
quartz are added along the NaAlSiO4-SiO2 join. In this system three more sodalite-forming
reactions can be generated (Sharp et al., 1989), i.e.
12 NaAlSi2O6+ 2 NaCl = Na8Al6Si6O24Cl2 + 6 NaAlSi3O8
jadeite
halite
sodalite
albite
6 NaAlSi2O6+ 2 NaCl = Na8Al6Si6O24Cl2 + 6 SiO2
jadeite
halite
sodalite
halite
sodalite
(12-3)
quartz
6 NaAlSi3O8+ 2 NaCl = Na8Al6Si6O24Cl2 + 12 SiO2
albite
(12-2)
(12-4)
quartz
Reaction (12-4) may be responsible for formation of sodalite in the Swartbooisdrif fenites.
However, Sharp et al. (1989) found the stability field of quartz + sodalite, limited by reactions
(12-3) and (12-4), to lie at temperatures below 298 K, and thus at geologically unreasonable low
temperatures. Conformably, natural equilibrium assemblages of sodalite + jadeite and sodalite +
12 Introduction
169
quartz have never been reported, as holds also true for the Swartbooisdrif fenite samples of this
study.
(2) The stability of sodalite under hydrothermal conditions has been studied by
Kostel`nikov & Zhornyak (1995), assuming the stability of sodalite to be defined by reaction
(12-2). However, in contrast to Sharp et al. (1989), the authors determined a stability field of
sodalite depending on temperature and salt concentrations in hydrothermal solutions. Starting
minerals were natural nepheline (with the potassium being removed by treatment with molten
NaCl), natural sodalite (hackmanite variety) and synthetic sodalite. The experiments were
performed at a constant pressure of 3 kbar and 600-800°C, with the starting minerals being
treated with fluids of variable salt concentrations of 10.0-42.5 wt.%. Only two phases were
found in the system, i.e. sodalite and nepheline, with the stability of sodalite with respect to
nepheline increasing with increasing NaCl-concentrations of the fluid. As temperatures rise,
more NaCl is needed to stabilise sodalite. Results of the experimental investigations of
Kostel`nikov & Zhornyak (1995) are shown in the T-X diagram of Fig. 12.3.
Sdl
NaCl (wt.%)
40
30
Sdl
+ Ne
Ne
Sdl
20
Ne + NaClaq
10
Pfluid = 3 kbar
600
700
800
Temperature (°C)
Fig. 12.3: Stability of sodalite in relation to temperature and NaCl concentration, determined by Kostel`nikov &
Zhornyak (1995).
However, both experimental investigations are of only restricted use for the
determination of the P-T conditions of the sodalite-formation in the studied samples, since
sodalite of the Swartbooisdrif area is clearly not an orthomagmatic crystallization phase, nor is it
exclusively formed by the hydrothermal alteration of nepheline (except in one nepheline syenite
sample), but represents a replacement product of plagioclase from fenitized anorthositic rocks
170
12 Introduction
incorporated by and bordering carbonatite centres (see Chapter 13). Moreover, quartz has never
been observed in these samples, hence suggesting sodalite formation under contemporaneous
desilication of the fenites. Unfortunately, stability relations for a conspicuous system like this are
still wanting.
13 Petrography
171
13 Petrography
13.1 Fenitized anorthosite
The magmatic petrography and mineral chemistry of the anorthositic rocks of the KIC was
discussed in detail in chapters 5.1 and 6. In summary, both, the white and the dark anorthosite
suite, display the primary magmatic assemblage of plagioclase (An37-53 and An43-75, respectively)
± olivine (XMg:0.54-0.65) ± orthopyroxene (XMg: 0.53-0.71) ± clinopyroxene (XMg: 0.66-0.76) ±
biotite (XMg: 0.39-0.67) + ilmenite + magnetite. In contrast to the dark anorthosite, the white
anorthosite is characterized by a pervasive alteration of all silicate phases, sericitization and
saussuritization of the feldspars and chloritization of the Mg-Fe silicates.
White anorthosite in contact to the carbonatitic breccia is of no great use for this study, as
the pervasive hydrothermal alteration of its primary magmatic mineral assemblage prevents a
clear distinction between late-magmatic hydrothermal processes and the subsequent
metasomatism by fenitizing fluids. Thus, this part of the study concentrates on the following
dark anorthosite samples:
Weakly fenitized pyroxene-bearing anorthosite: Ku-98-78
Weakly fenitized leucogabbronorite: Ku-98-71, Ku-98-72, Ku-98-79, Ku-98-84
Weakly fenitized leucotroctolite: Ku-98-77
Fenitized anorthosite: Ku-98-19, Ku-98-36, Ku-98-54, Ku-98-114, Ku-98-120, Ku-99-01, Ku-99-02
Fenitized anorthosite xenoliths in the CB: Ku-98-55, Ku-98-89
Sample Ku-98-77, -78, -79 and -84 were collected at a distance of c. 2.0, 2.5, 2.5, and <0.1 m
from the contact with a dyke of the carbonatitic breccia, respectively, in an outcrop located about
2 km N of the operating quarry. Samples Ku-98-71 and –72 are sampled in 20 cm and 2 m
distance from the contact with another dyke. Samples Ku-98-19, -36, -54, -114, -120 and Ku-9901 and -02 were taken from direct contact zones between anorthosite and dykes of the
carbonatitic breccia, where the fenitizing processes reached a maximum magnitude.
When approaching the contact to dykes of the carbonatitic breccia, plagioclase (An49-54)
of the dark anorthosite suite still shows oscillatory zoning but is progressively altered to
sericite/muscovite and albite to oligoclase (An1-17; Fig. 13.1a). Clinopyroxene, orthopyroxene
and/or olivine are increasingly replaced by assemblages of deep green biotite (Fig. 13.1b),
172
13 Petrography
Fig. 13.1: Polished-section photomicrographs illustrating the mineralogical changes of anorthositic rocks during
progressive fenitization. Cross-polarized light (a, e-f). Plane-polarized light (b-d). a) Replacement of orthomagmatic
plagioclase by albite. b) Alteration of pyroxene to biotite. c) Replacement of clinopyroxene by fine-grained calcitemagnesio-riebeckite assemblages. d) Anhedral fluorite grains filling cracks in plagioclase. e) Sodalite replacing
plagioclase along both its grain margins and albite-twin lamellae. f) Sodalite and ankerite replacing plagioclase
along its twin lamellae.
13 Petrography
173
subhedral epidote and granular calcite, chlorite-calcite intergrowths, and/or fine-grained
intergrowths of bluish magnesio-riebeckite and calcite (Fig. 13.1c). The amount of apatite,
associated with biotite and calcite/ankerite, strongly increases in the same direction. As is
suggested by the brecciation of plagioclase, the deformation of Fe-Ti oxides and the increasing
abundance of cracks in the rocks, the investigated dark anorthosite samples underwent an intense
deformation prior to, or contemporaneous with the carbonatite emplacement. Joints in the rocks
are commonly filled by epidote-calcite and magnesio-riebeckite-calcite intergrowths (Ku-98-71).
In one sample (Ku-98-89), fissures in the altered plagioclase are filled by conspicuous anhedral
fluorite grains, displaying a patchy colour zonation from light blue to lilac (Fig. 13.1d).
At the direct contact to large dykes of the carbonatitic breccia, the dark anorthosites are
almost completely transformed into purplish-blue coloured anorthositic fenites, with the
magmatic precursor mineralogy and textures being obliterated by the fenitizing processes. The
rocks generally exhibit a banded appearance due to an alternation of albite (An0-3)/sodalite -,
ankerite- and biotite-rich layers.
Plagioclase of the fenites is commonly heavily strained. Most of the coarse-grained
crystals are broken down to fine-grained angular particles, displaying undulose extinction. If
preserved, the primary-magmatic cumulus plagioclase is completely replaced by almost pure
albite (An0-3), that exhibits bent and diffuse albite twin lamellae and may be associated with
minor ankerite. Locally, plagioclase is recrystallised to mosaics of granular, twinned to
untwinned albite (An0). Poikilitic secondary albites have been observed, enclosing abundant
biotite, muscovite and carbonate inclusions, hence suggesting that biotite and muscovite
formation as well as carbonate injection was synchronous with the plagioclase recrystallisation.
Sodalite, causing the purplish macroscopic colour of the anorthositic fenite, is restricted to the
albite-rich zones, where it was formed at the expense of plagioclase. Various styles of
replacement have been observed, i.e. (1) sodalite replacing the marginal zones of relict cumulus
plagioclase (Fig 13.1e), (2) sodalite exclusively grown at the expense of the albite twin lamellae
of relict cumulus plagioclase (Fig. 13.1f), (3) sodalite pseudomorphing the shape of secondary,
granular albite and (4) sodalite replacing accumulations of broken plagioclase particles, forming
almost monomineralic, diffuse schlieren (Fig. 13.2a). In addition, sodalite forms narrow rims
around biotite-ankerite accumulations and magnetite. These almost monomineralic rims are
interpreted in terms of an elevated fluid activity, that mainly affected the weak contacts and
interstices between individual plagioclase crystals. Sodalite of all textural positions bears
inclusions of biotite, albite and/or ankerite, giving the grains a poikilitic appearance (Fig. 13.2b).
The fact that plagioclase relics, preserved in sodalite are always pure albite is taken as evidence
174
13 Petrography
that the transformation of Ca-rich, primary-magmatic plagioclase into albite predates the sodalite
formation. Thus, sodalitization of plagioclase is at least a two-stage process, according to the
following simplified reactions:
CaNa[Al1.5Si2.5O8] + 0.5 Na+ + 2 Si4+ = 1.5 Na[AlSi3O8] +0.5 Al3+ + Ca2+
intermediate Pl
in solution
Ab
in solution
6 Na[AlSi3O8] + 2 NaCl = Na8[Al6Si6O24]Cl2 + 12 SiO2
Ab
in solution
Sdl
(13-1)
(13-2)
in solution
Albitization of plagioclase following reaction (13-1) could thus have produced part of the Ca,
needed for the formation of the associated ankerite. Remarkably, this reaction is in need for a
certain amount of Si, which is a rather unlikely species for carbonatite-derived fluids, thus
suggesting that either (1) the Si necessary for the reaction was produced by additional, Sireleasing reactions or (2) that both Si and Al behaved mobile during metasomatism, leading to
misfits when balancing for Al. In clear contrast, reaction (13-2) is in excellent agreement with
fenitization, caused by Si-undersaturated, NaCl-rich fluids as could be released by carbonatite
melts.
Greenish to brownish biotite is mostly recrystallised to polygonal grains, forming
irregularly bounded clusters (Fig. 13.2c). However, even the recrystallised grains display
undulose extinction, pointing to multiple deformation events. The mica is commonly associated
with minor ankerite and fine-grained magnetite, with the latter being expulsed from biotite
during the recrystallisation event. The fact, that biotite is not evenly distributed within single thin
sections but forms mafic enclaves, similar in shape and distribution to those of unaffected
anorthosites, is taken as evidence that biotite was mainly formed at the expense of magmatic
accumulations of intercumulus clinopyroxene, orthopyroxene, olivine and/or late-magmatic
amphibole. However none of these minerals nor their alteration products have been observed in
the investigated samples, hence suggesting that the fenitization of these rocks reached a
pervasive extent.
Granular ankerite, an abundant phase in the anorthositic fenite, is commonly associated
with biotite, displaying straight grain boundaries against the latter (Fig. 3.2.d). Recrystallisation
of ankerite to granular mosaics is a common feature. In addition, ankerite is enriched in almost
monomineralic, irregularly bounded schlieren and fills small cracks in the rock.
Predominant ore minerals of all samples investigated are magnetite and ilmenite. Many of
the magnetites display oxidative exsolution of ilmenite lamellae //{111} and {100}, similar in
composition to larger ilmenite grains. Along the two lamellar systems and along cracks,
13 Petrography
175
Fig. 13.2: Polished-section (a-d) and ore section (e-f) photomicrographs illustrating the mineralogical changes of
anorthositic rocks during progressive fenitization. Cross-polarized light (a). Plane-polarized light (b-d). a) Sodalite
forming diffuse schlieren, replacing fragmented plagioclase particles. b) Poikilitic sodalite, bearing abundant
inclusions of ankerite and biotite. c) Heavily strained biotite, recrystallised to polygonal grains. d) Anhedral,
granular ankerite, displaying straight grain boundaries against biotite. e) Replacement of magnetite by ankerite,
whereby ilmenite exsolution lamellae are preserved. f) Isolated pyrite grains, occurring in the vicinity of ilmeniteexsolution lamellae.
176
13 Petrography
magnetite is partially altered to hematite, which is in turn replaced by ankerite, whereby ilmenite
exsolution lamellae are mostly preserved (Fig. 3.2e, f). Along ilmenite lamellae, small, isolated,
anhedral grains of pyrite have formed (Fig. 3.2e, f), whereas cracks in larger ilmenite grains are
often filled with massive pyrite aggregates. In addition, ilmenite may be surrounded by narrow
rutile rims. In sample Ku-98-79, a large pyrite grain is overgrown by pyrite of a second
generation. The former {100} grain boundary of the pyrite crystal is marked by trails of fluid
inclusions. The amount of sulphides as well as the intensity of carbonate replacement of the
oxides increase towards the contact with the carbonatite dykes. In sample Ku-98-71, shear bands
in oxides and δ clasts of ilmenite testify to strong deformation at the dyke contacts. The observed
reaction relationship of the Fe-Ti oxides with ankerite and pyrite suggests that carbonitization
and sulphidation took place due to the interaction of ilmenite and magnetite with O2, S2 and CO2
in the fenitizing fluid phase. Carbonatization of the Fe-Ti oxides produced rutile, hematite and
the siderite-component in ankerite following the simplified reactions (Yang, 1987)
FeFe2O4 + CO2 = Fe2O3 + FeCO3
(13-3)
FeTiO3 + CO2 = TiO2 + FeCO3
(13-4)
Sulphidation of Fe-Ti oxides appears to be a multiple event, which is obviously related to the
metasomatic overprint. Sulphidation of ilmenite and magnetite, according to the reactions
(Barton and Skinner, 1979)
FeFe2O4 + 3 S2 = 3 FeS2 + 2 O2
(13-5)
4 FeTiO3 + S2 = FeS2 + FeFe2O4 + 4 TiO2
(13-6)
takes place in most anorthosite samples in contact with dykes of the carbonatitic breccia. A twostage replacement is clearly indicated by the pyrite overgrowths on a larger pyrite crystal in the
anorthosite sample Ku-98-79. Since no early-magmatic sulphide phases were recorded in the
KIC rock types investigated, the sulphidation must be related to the repeated influx of S-rich
fenitizing fluids, presumably released by the intruding carbonatite melts.
The absence of anhydrous Fe-Mg silicates and the progressive transformation of calcic
plagioclase into pure albite and sodalite provides evidence for the hydrous and alkali-rich nature
of the fenitizing fluid. The extensive formation of biotite, suggests that these fluids additionally
contained a certain amount of K. Judging from the total of newly grown minerals, formed during
13 Petrography
177
metasomatism, further species of the fluid must include certain amounts of CO2, P, Cl, F, S and
Fe3+.
13.2 Fenitized syenite
Syenite unaffected by metasomatism (see Chapter 5.2 and Chapter 6) is characterized by a
homogeneous texture with coarse to medium-grained K-feldspar crystals without preferred
orientation. The main minerals present are perthitic K-feldspar (orthoclase/microcline) and
plagioclase (An0-18). Late magmatic quartz, clinopyroxene (XMg: 0.45-0.51) and hastingsite (XMg:
0.01-0.06) may be present in varying amounts, forming discrete grains in the interstices between
feldspar crystals. Common accessories include titanite, chlorite, epidote, Fe-Ti oxides and
zircon. For the investigation of the fenitizing processes, affecting syenites emplaced in shear
zones subsequently intruded by carbonatites, the following 16 samples have been studied in
detail:
Fenitized syenite:
Ku-97-05a, Ku-98-59, Ku-98-66, Ku-98-69, Ku-98-70s, Ku-98-73s,
Ku-98-74, Ku-98-93, Ku-98-103s, Ku-98-105, Ku-98-106, Ku-98107, Ku-99-06, Ku-99-09, Ku-99-10, Ku-99-18, Ku-99-19s
When approaching the carbonatitic breccia, the grain size of syenite decreases from
medium- to fine-grained, following shearing and recrystallisation previous to and outlasting the
emplacement of the carbonatites. The main constituent of the granular matrix is mostly
untwinned and subordinately twinned and subrounded albite (An0-5) with undulose extinction
(Fig. 13.3 a, b), together with minor aggregates of recrystallized biotite (Fig. 13.3 a) and/or finegrained, isometric muscovite (Fig. 13.3b), presumably formed at the expense of former Kfeldspar. Fragmented relics of fine-grained microcline, which are apparently completely replaced
by chess-board albite (Fig. 13.3c), support this interpretation, suggesting K-feldspar replacement
according to the simplified reaction:
K[AlSi3O8] + Na+ = Na[AlSi3O8] + K+
Kfs
(13-7)
Ab
providing the K needed for the simultaneous muscovite formation. The alteration of K-feldspar
under simultaneous formation of equal amounts of albite and muscovite can be written as Alconservative reaction:
178
13 Petrography
Fig. 13.3: Polished-section photomicrographs illustrating the mineralogical changes of syenites during progressive
fenitization. Cross-polarized light (a-d, f). Plane-polarized light (e). a) Fine-grained, granular albite, displaying
undulose extinction, is the main constituent of the fenitized syenites, associated by minor and strongly recrystallized
biotite. b) Albite associated by fine-grained, isometric muscovite. c) Relics of microcline, replaced by chess-board
albite are preserved locally. d) Granular ankerite, muscovite and albite, forming part of the even grained matrix. e)
Magnetite and biotite, accumulated in irregular streaks. f) Micro-shearzones in the fenitized syenites testify to a
strong deformation of the rocks.
13 Petrography
179
4 K[AlSi3O8] + Na+ + 2 H+ = Na[AlSi3O8] + K Al2[(OH)2| AlSi3O10] + 6 SiO2 + 3K+
Kfs
in solution
Ab
Msc
(13-8)
in solution
which, in contrast to reaction (13-7), additionally accounts for the generally Si-deficient nature
of carbonatite-derived fluids. The lack of dust-like Fe oxide inclusions (magnetite/hematite) in
the secondary albite of the fenitized syenite is responsible for the white to pale pink colour of
these rocks.
Fine-grained and granular ankerite is evenly distributed in the matrix, intimately
associated by albite and muscovite (Fig. 13.3d). Minor heavily strained and partially
recrystallised biotite, as well as anhedral to subhedral magnetite and ilmenite grains occur as
patches or as intensely deformed irregular streaks (Fig. 13.3e). Pyrite, chlorite, epidote and
apatite are common accessories. The amount of zircon generally increases with progressive
fenitization; locally, zircon of a second generation forms rims around presumably magmatic
zircon cores. Micro-shearzones (Fig. 13.3f) and repeated recrystallisation testify to several
deformation events during the subsequent injections of carbonatite melts and/or repeated
movements along the hosting faults. Fissures in the rock are invaded by magnetite and ankerite.
Hastingsite or clinopyroxene relics or their alteration products have never been observed
in the investigated samples, suggesting that the syenites were pervasively altered by the
circulating hydrous fluids. Remarkably, however, sodalite replacing secondary albite of the
fenitized syenites is virtually completely absent (except one sample, Ku-98-59b). A possible
explanation for this lack of a second stage of transformation is that the highly mobile, NaCl
saturated fluids migrated in the anorthositic lithologies bordering the shear zones, thus leaving
back fluids which contained insufficient NaCl to form sodalite. Accessory, very fine-grained
analcite, replacing albite, has been observed in one sample (Ku-98-59b)
Contacts between the syenites and the carbonatitic breccia are always sharp. Syenite
fragments in all stages of detachment can be observed in the carbonatitic breccia, hence
constraining an intrusive relationship between the two units, with the younger carbonatites
crosscutting the previously emplaced syenites.
13.3 Fenitized nepheline syenite
Nepheline syenite (Ku-99-15a, -15b, -15c and –15d) was observed at only one locality of the
mapped area, where the small and impersistent dyke borders the carbonatitic breccia. The rock is
180
13 Petrography
Fig. 13.4: Polished-section photomicrographs illustrating the mineralogical changes of the nepheline syenite during
progressive fenitization. Cross-polarized light (a-b, d-f). Plane-polarized light (c). a) Contact between marginally
altered nepheline and K-feldspar, the latter being unaffected by the fenitization. b) Medium-grained, perthitic Kfeldspar laths. c) Anhedral and heavily strained biotite, filling the interstices between individual K-feldspar crystals.
d) Cancrinite-muscovite-sodalite assemblage, replacing nepheline. e) Anhedral ankerite and muscovite, filling
cracks in nepheline. f) Sodalite, replacing nepheline along cracks.
13 Petrography
181
Fig. 13.5: Polished-section photomicrographs of the carbonatitic breccia and an enclosed ultramafic xenolith, that
underwent extensive fenitization. Cross-polarized light (a, c-e). Plane-polarized light (b, f). a) Subhedral dolomite,
as well as anhedral calcite, biotite and magnetite are the main constituents of the enclosing carbonatitic breccia. b)
Anorthositic wallrock-xenoliths in the carbonatitic breccia are mainly composed of albite, with the latter being
partially replaced by sodalite and ankerite. c) Fine-grained magnesite, magnesio-arfvedsonite and magnetite are the
major constituents of the fine-grained matrix of the ultramafic xenoliths. d) Magnesite, magnetite and magnesioarfvedsonite, forming veins. e) The contacts between the ultramafic xenoliths and the enclosing carbonatitic breccia
are marked by two-fold corona structures, being composed of biotite and ankerite. f) Anhedral, poikilitic aegirine
grain.
182
13 Petrography
characterized by a streaky appearance due to changes in the grain sizes of its major phases
nepheline and K-feldspar from medium- to coarse-grained.
Rare unaffected parts of nepheline syenite show the medium- to coarse-grained euhedral
nepheline (Fig. 13.4 a) and subhedral tabular to lath-shaped, mesoperthitic K-feldspar (Fig. 13.4
a, b) to be the cumulus phases of the rock. Minor post-cumulus biotite, filling the interstices
between individual nepheline and K-feldspar crystals (Fig. 13.4c), is the dominant mineral
among mafic silicates. It occurs as small anhedral and heavily strained grains. Accessory, late
stage epidote is localised in patches.
When reaching the ferrocarbonatite contacts, the grain size of nepheline attains maximum
diameters of 15 cm, suggesting that the carbonatite emplacement triggered the nepheline growth.
Muscovite and cancrinite are mainly localised at triple junctions between individual nepheline
grains (Fig. 13.4 d), but may also occur in small fissures transsecting nepheline (Fig. 13.4 e). The
textural relationships strongly suggest that both muscovite and cancrinite were formed at the
expense of nepheline, following the injection of fenitizing fluids. Granular ankerite fills cracks in
the rock (Fig. 13.4e). Along its margins and irregular cracks, nepheline is replaced by sodalite
(Fig. 13.4f), according to reaction
1.5 Na3[AlSiO4]4 + 3.5 Na+ + 2 Cl- = Na8[Al6Si6O24]Cl2
Ne
in solution
(13-9)
Sdl
Reaction (13-9) implies an introduction of Na and Cl to the system and thus agrees well with the
interpretation, that NaCl-rich, carbonatite-derived fluids migrated through country rocks
neighbouring the carbonatite centres. Moreover, the replacement of nepheline by sodalite
accounts for the macroscopically observed colour change of nepheline from pale yellow to pale
blue when reaching the contacts to the carbonatite.
13.4 Fenitized ultramafic xenoliths
Ultramafic rocks (Ku-97-43a, Ku-97-43b) occur as angular xenoliths within one large dyke of
the sodalite-rich carbonatitic breccia.
13 Petrography
183
The matrix of the enclosing carbonatitic breccia is mainly composed of euhedral to
subhedral ankerite/dolomite, interstitial calcite and anhedral magnetite grains (Fig. 13.5 a).
Abundant anorthositic wallrock fragments are surrounded by the carbonatite matrix, and mainly
consist of albite, being largely replaced by sodalite and ankerite (Fig. 13.5 b), and minor, heavily
strained biotite.
Subhedral magnesite (53-68 vol.%), bluish to pale bluish magnesio-arfvedsonite (30-45
vol.%) and granular magnetite (~ 5 vol.%) are the major constituents of the fine-grained,
equigranular matrix of the fenitized ultramafic xenoliths whereas ankerite is present in minor
proportions (Fig. 13.5c). Granular magnesite and subhedral to euhedral, lath-shaped magnesioarfvedsonite may also form fine to medium-grained veins and schlieren (Fig. 13.5d). Mediumgrained biotite and ankerite occur localised in patches and, more common, as two-fold corona
structures at the contacts between the xenoliths and the enveloping carbonatite (Fig. 13.5e). In
the latter textural position biotite together with minor magnetite marks the outer contact zones,
whereas the inner parts of the corona are almost completely composed of ankerite. Fine-grained,
anhedral aegirine grains of up to 1.2 mm in diameter have been observed in the magnesitemagnesio-arfvedsonite-magnetite matrix. The aegrines are always strongly poikilitic, bearing
abundant magnesite, magnesio-arfvedsonite and ankerite inclusions (Fig. 13.5f).
Magnesite was most probably formed by replacement of former Fe-Mg silicates
(presumably olivine), caused by the injection of CO2-rich fluids. Such a fenitization process
would be in excellent accordance with an interaction between the ultramafic xenoliths and the
volatile-rich carbonatite melt which brought them to the surface.
13.5 Carbonatitic breccia
Since the composition and internal structure of the carbonatitic breccia are intricate and change
on a millimetre scale, only a simplified description of this rock type can be given. Predominant
minerals are albite, ankerite, calcite, biotite, sodalite, muscovite and Fe-Ti oxides. Apatite,
cancrinite, pyrochlore and sulphides are rare. A detailed description of the sulphide- and oxidemineralogy and textures of this rock-type has been presented by von Seckendorff & Drüppel
(1999) and von Seckendorff et al. (2000). Based on field observations the carbonatitic breccia
has been subdivided into three units, i.e. (1) massive carbonatitic breccia, (2) layered carbonatitic
breccia and (3) sodalite-rich carbonatitic breccia (see Chapter 3.3.6.1). Investigated samples of
these units include:
184
13 Petrography
Massive carbonatitic breccia:
Ku-98-08, Ku-98-12, Ku-98-63, Ku-98-73c, Ku-98-103c, Ku-
98-104, Ku-98-105, Ku-98-106
Layered carbonatitic breccia :
Ku-98-05, Ku-98-14, Ku-98-28, Ku-98-48, Ku-98-51, Ku-98-
55, Ku-98-56, Ku-98-62, Ku-98-64, Ku-98-69, Ku-98-70c, Ku-98-87, Ku-98-89, Ku-98-118,
Ku-99-04, Ku-99-17, Ku-99-18, Ku-99-19c
Sodalite-rich carbonatitic breccia: Ku-97-15, -15a, -15b, -15c, -15d, -15e, -15g, -15i, -15k, 15l, 15n, -15o, -15q, -15r, -15s, -15t, -15u, -15v, Ku-97-19, Ku-97-24, Ku-97-26, Ku-98-17, Ku-9818, Ku-98-27, Ku-98-31, Ku-98-47, Ku-98-57a, Ku-98-57b, Ku-98-58a, Ku-98-58b, Ku-98-80,
Ku-98-83, Ku-98-131, Ku-99-05, Ku-99-06, Ku-99-07, Ku-99-08, Ku-99-SA8, Ku-99-SA11,
Ku-01-OD1, Ku-01-OD2, Ku-01-OD3
REE-rich samples: Ku-98-130a and Ku-98-130b
13.5.1 Massive carbonatitic breccia
This sub-unit of the carbonatitic breccia is characterized by a macroscopically massive
appearance. However, on a microscopic scale, almost all samples investigated display a narrow
lamination and small-scaled magmatic folding (Fig. 13.6a). The erroneously macroscopic
impression of textural homogeneity of the massive carbonatitic breccia is caused by the small
grain sizes of the heavily brecciated and partially recrystallised minerals in this rock type, which
commonly range between 1-30 µm and rarely exceed 0.5 mm.
The main minerals present are anhedral ankerite (50-75 vol.%) and subrounded fragments
of albite (10-25 vol.%). Locally, larger relics of plagioclase with diameters of up to 500 µm have
been observed. The bent and fissured, anhedral grains are surrounded by the laminated ankeritealbite matrix, giving the rock a mylonite-like appearance on a microscopic scale (Fig. 13.6 b).
Subordinate fragmented magnetite is dispersed in the groundmass. Rare sodalite replaces
schlieren of fragmented plagioclase. Other phases are present in only minor amounts (< 3 vol.%)
and too small to be identified either microscopically or with the help of EMP or XRD analysis.
13 Petrography
185
Fig. 13.6: Polished-section photomicrographs of the massive carbonatitic breccia (a-b) and the layered carbonatitic
breccia (c-f). Cross-polarized light (c-d). Plane-polarized light (a-b, e-f). a) Small-scaled lamination and magmatic
folding of the massive carbonatitic breccia. b) Very fine-grained, laminated carbonatite, surrounding a rotated
wallrock xenolith. c) Layer composed of granular ankerite, magnetite and twinned albite. d) Layer composed of
heavily strained albite fragments, displaying undulose extinction and being partially replaced by muscovite. e) Layer
composed of biotite, magnetite, albite, and euhedral pyrochlore. f) Layer composed of orthomagmatic apatite,
biotite, magnetite, and minor ankerite.
186
13 Petrography
13.5.2 Layered carbonatitic breccia
Samples of the layered carbonatitic breccia display an indistinct layering and magmatic folding
on a mm to cm scale, with extremely variable grain sizes ranging from fine- to coarse-grained in
different layers. Single layers are commonly composed of (1) irregularly bounded ankerite
grains, partially displaying undulose extinction (65-80 vol.%), associated by oval-shaped and
twinned albite (18-35 vol.%), bearing carbonate and magnetite inclusions, rare schlieren of
sodalite and dispersed, granular magnetite (Fig. 13.6c), (2) fragmented anhedral to subrounded
albite, displaying patchy, undulose extinction (Fig. 13.6d), (3) partially recrystallised greenish
biotite (37-50 vol.%), together with minor anhedral magnetite, anhedral, embayed albite grains
(45-57 vol.%) and accessory pyrochlore (Fig. 13.6e) and (4) greenish biotite, subhedral to
euhedral apatite prisms, granular ankerite and minor magnetite (Fig. 13.6f). Sodalite is a minor
constituent of these rocks, even though it locally forms diffuse schlieren, pseudomorphing
fragmented albite. Monomineralic cancrinite laminae are mainly restricted to the contact zones
between carbonate- and feldspar-rich layers and are thus suggested to be formed by metasomatic
exchange between the incorporated feldspar-rich xenoliths and the carbonatite melt. However,
the intermingling of carbonatitic and xenolithic material in a first step and the subsequent
separation of minerals into individual layers make it difficult to unravel the exact temporally
succession of processes.
13.5.3 Sodalite-rich carbonatitic breccia
13.5.3.1 REE-poor samples
The mostly medium to coarse grain size of the sodalite-rich carbonatitic breccia, makes the
separation between the crystallized carbonatite on the one hand and xenolithic material on the
other hand a lot easier. In the samples investigated, subhedral to anhedral and partially
recrystallised granular ankerite (5-40 vol.%) is the main constituent of the cementing
ferrocarbonatite matrix, whereas subordinate anhedral calcite fills the interstices between
individual ankerite grains (Fig. 13.7a). Two generations of ankerite can be distinguished: (1) a
first generation of subhedral and Mg-rich ankerite is preserved in rare cases (Fig. 13.5 a), and (2)
a second generation of anhedral, granular and locally recrystallised ankerite (Fig. 13.7 a, b),
differing from the former in its distinctly higher Fe-contents. In addition, small grains of Mn-rich
ankerite were observed in the vicinity of fragmented Fe-Ti oxide grains.
13 Petrography
187
Fig. 13.7: Polished-section photomicrographs of carbonatite minerals from the sodalite-rich carbonatitic breccia.
Cross-polarized light (a-b). Plane-polarized light (c-f). a) Anhedral, granular ankerite, associated by minor
magnetite, forming an even grained matrix. b) Ankerite, marginally recrystallised to polygonal grains. c) Apatite,
bearing abundant inclusions of albite, magnetite and ankerite. d) Straight grain boundaries are developed between
early crystallized apatite, biotite and magnetite. e) Marginally, both biotite and apatite are recrystallised to polygonal
grains. f) Fine-grained pyrochlore, altered to very fine-grained intergrowths of secondary minerals.
188
13 Petrography
Accumulations of optically zoned apatite grains have been observed in two samples of
the sodalite-rich carbonatitic breccia. The euhedral to subhedral apatite prisms frequently contain
inclusions of ankerite, magnetite and minor albite (Fig. 13.7 c, d). Rarely, biotite is included in
large apatite grains (Fig. 13.7 c). Apatite is intimately associated with, and displays straight grain
boundaries against subhedral biotite and a first generation of euhedral to anhedral and almost
inclusion-free magnetite of up to 0.5 mm in diameter (Fig. 13.7d). Apatite, biotite and magnetite
appear to have co-crystallized early in the magmatic history of the carbonatite, even though
biotite and magnetite have overgrown apatite locally and thus outlasted the apatite formation. At
their margins, both biotite and apatite are partially recrystallised to granular mosaics, supporting
evidence that shearing outlasted the carbonatite emplacement and crystallization (Fig. 13.7 e).
Small, unzoned pyrochlore crystals of up to 100 µm in diameter are common accessories of the
cementing carbonatite matrix (Fig. 13.7 f). However these crystals are virtually completely
replaced by fine-grained intergrowths of alteration products, which, however, where too small to
be unambiguously identified.
Layers enriched in relict fenitized wallrock material are mainly composed of irregularly
bounded fragments of fine-grained to very fine-grained albite displaying diffuse undulose
extinction and containing abundant biotite inclusions (Fig. 13.8 a). In layers composed of both
ankerite and albite, the latter has the tendency to develop a subrounded shape, suggesting
solution of the albite grain margins by the carbonatite magma. In clear contrast to the angular
albite fragments, these albites contain abundant ankerite, biotite and magnetite inclusions, mostly
lack undulose extinction and may or may not display albite twinning (Fig. 13.8 b). In cases
where twinning is missing, these feldspars strongly resemble quartz. In addition accumulations
of twinned to untwinned albite (An0-1; 25-90 vol.%), in places associated with brownish to
greenish, heavily strained biotite and fine-grained magnetite, form irregularly bounded zones
pseudomorphing former xenolith boundaries. The latter are rotated and surrounded by narrow,
carbonate-rich laminae.
The replacement of albite by sodalite started during the end-stages of biotite formation as
is evidenced by biotite-sodalite inclusion-relationships. Sodalite occurs in different textural
positions, i.e. (1) as virtually monomineralic layers, where sodalite replaces very fine-grained
fragmented albite or relics of chess-board albite (Fig. 13.8 c), (2) as a replacement product of
brecciated and albite-rich wallrock xenoliths, where the feldspathoid may be accompanied by
ankerite (Fig. 13.8 d) and surrounded by narrow reaction rims of subhedral biotite and/or small
euhedral magnetite grains and (3) as almost monomineralic rims around biotite-apatite patches
and magnetite. Sodalite is mostly poikilitic, bearing abundant inclusions
13 Petrography
189
Fig. 13.8: Polished-section photomicrographs of the sodalite-rich carbonatitic breccia. Cross-polarized light (a-d).
Plane-polarized light (e-f). a) Fragmented albite, displaying undulose extinction and containing abundant biotite
inclusions, is the main constituent of the incorporated wall-rock xenoliths. b) In layers predominantly composed of
ankerite, twinned to untwinned albite displays a rather oval shape. c) Sodalite occurring in schlieren, replacing
fragmented albite. d) Sodalite replacing rotated, albite-rich anorthositic xenoliths. e) A yet unspecified Na-Al phase,
displaying a fibrous morphology, replaces sodalite. f) Metasomatically formed biotite occurs in almost
monomineralic schlieren, oriented parallel to the overall texture of the rock.
190
13 Petrography
Fig. 13.9: Polished-section (a-e) and ore section (f) photomicrographs of the carbonatitic breccia. Cross-polarized
light (a-e). a) Granular albite, either replaced by muscovite or sodalite. b) Muscovite replacing albite, whereby early
formed sodalite is preserved. c) At the contacts between sodalite and carbonate-rich layers, narrow monomineralic
cancrinite rims are developed. d) Late-stage cancrinite, replacing sodalitized albite. e) Late-stage, poikilitic calcite,
bearing abundant inclusions of the matrix minerals. f) Millerite lamellae in thiospinel of the polydymite-violarite
group, replaced by chalcopyrite.
13 Petrography
191
of fine-grained ankerite and magnetite. Along both its margins and cracks sodalite is altered to a
yet unknown silicate-free sodium-aluminium phase (Drüppel et al., in prep.). This phase
commonly forms bunches of concentrically radiating fibres, with individual crystallites being up
to 30 µm in length and < 1 µm in width (Fig. 13.8 e).
Dark green biotite occurs in sodalite-rich and more commonly in albite-magnetite layers
and patches, where the anhedral, embayed mica grains are heavily stained. In addition, biotite
has been observed in carbonate-rich zones, where it forms schlieren of recrystallised grains,
elongated parallel to the layering of the rocks (Fig. 13.8 f). Biotite of both textural positions is
interpreted to be a relict phase of the incorporated anorthositic xenoliths.
Subordinate isometric muscovite and cancrinite locally replace fine-grained, fragmented
albite, whereby early formed sodalite is preserved (Fig. 13.9 a, b, d). Cancrinite may also be
present as reaction rims around albite- and sodalite-rich patches, separating carbonate-rich from
silicate-rich zones (Fig. 13.9 c). Both minerals are rarely included in sodalite but are commonly
enclosed by magnetite of a second generation. Late-stage calcite occurs in most of the samples,
where it overgrows all previously mentioned minerals (Fig. 13.9 e). Chlorite and epidote are
common accessories.
Magnetite is the dominant mineral among Fe-Ti oxides in the sodalite-rich carbonatitic
breccia. A first generation of almost inclusion-free magnetite has been observed in carbonaterich zones, where the oxide occurs as a widespread fine-grained matrix mineral, but may also
form larger grains of up to 4 mm in size. Individual grains are subhedral to anhedral and display
straight, curved or embayed grain boundaries with ankerite, biotite and apatite. The second
generation of magnetite may reach grain sizes of up to 2 cm in diameter, frequently displaying
inclusions of albite, sodalite, apatite, biotite and ankerite as well as rare muscovite and cancrinite
(Fig. 13.9e). In either case, most of the magnetite grains are anhedral with undulating grain
boundaries, embayments and cavities, whereas subhedral to euhedral grains are rare. Many of the
magnetite grains are rotated and fragmented, the fractures being filled with carbonate or pyrite.
Additionally, magnetite has been observed as vein fillings, where it is intergrown with
titanohematite. The coexistence of hematite and magnetite points to T-f(O2) conditions at the
hematite-magnetite buffer. Martitisation was observed in various stages of progression, in some
cases advanced to the point where the texture consists of larger hematite areas containing relict
magnetite. In addition, hematite forms fine-grained, anhedral or amoeboidal overgrowths on
magnetite, preferentially at grain contacts with ankerite. Subhedral ilmenite forms either single
grains or has developed several deformed subgrains. Lamellar twins after {10-11} are common
but often bent due to post-crystalline deformation. In irregularly bounded zones, ilmenite is
192
13 Petrography
altered to very fine-grained symplectitic aggregates of magnetite and rutile. Titanohematite was
found as an alteration product of ilmenite, invading the latter along cracks. Rutile is present in
various textures, i.e. (1) as large, anhedral grains, either isolated or in contact with magnetite, (2)
as groups of small, nearly euhedral to subhedral grains, as fine-grained overgrowths on
magnetite where it is associated with ankerite and (3) as lamellae in magnetite, where it may be
intergrown with hematite.
Pyrite is the dominant sulphide mineral. A first generation of pyrite occurs as subhedral to
euhedral grains, either (1) dispersed in the matrix, (2) in the vicinity of magnetite or (3) as
inclusions in chalcopyrite. Narrow fractures in the rocks are filled with a later generation of late
pyrite. Minor amounts of chalcopyrite, millerite, siegenite and violarite-polydymite were
identified, testifying to different cooling stages. The sulphides are frequently concentrated in
irregular sulphide veins or patches and elongate lenses where they may be associated with minor
magnetite and pyrite, but they may also form isolated grains in the matrix. Individual millerite
grains underwent partial or total replacement by thiospinels of the violarite-polydymite group
(Fig. 13.9 f). Sulphides, represented by members of the chalcocite-digenite group, associated
with rare chalcopyrite and bornite, were observed only in one sample (Ku-97-15d), where they
form small grains or grain aggregates, which are either dispersed in the matrix or, more
frequently, form inclusions in magnetite. Pyrrhotite may be present in small amounts (Ku-97-19)
forming single grains and irregular aggregates, concentrated in layer-parallel patches. In larger
pyrrhotite grains two systems of pentlandite lamellae may be developed. Conspicuous oxidesulphide aggregates have been observed locally, which have been investigated in some detail by
von Seckendorff et al. (2000). The inner parts of the aggregates mainly consist of granular
intergrowths of sub- to euhedral pyrite with anhedral magnetite, chalcopyrite and rare millerite.
The broad outer zones of the domains are characterized by lamellar intergrowths of pyrite,
hematite and magnetite with rare chalcopyrite, millerite and polydymite. In the outermost
regions of these domains, pyrite lamellae are replaced by millerite, together with polydymite and
chalcopyrite. These conspicuous aggregates have been interpreted to be formed by
contemporaneous sulphidation and oxidation of a suspected precursor pyrrhotite (see von
Seckendorff et al., 2000, for a detailed discussion).
13.5.3.2 REE-rich samples
The two REE-rich samples of the sodalite-rich carbonatitic breccia (Ku-98-130a, Ku-98130b) are characterized by a pale pink macroscopic colour (except abundant deep blue sodalite
fragments). Ankerite occurs as large anhedral grains with embayed grain boundaries, intimately
13 Petrography
193
associated with magnetite. The magnetite grains are generally anhedral with undulating grain
boundaries, embayments and cavities. Albite is the main constituent of fragmented xenoliths. In
these zones, however, albite is only rarely preserved, since most of the grains are pervasively
altered to fine-grained intergrowths of ankerite and sodalite, with the latter in turn being replaced
by fibrous bunches of the Na-Al phase (Fig. 13.10 a).
Fig. 13.10: Polished-section photomicrographs of the REE-rich samples of the carbonatitic breccia. a) Albite altered
to sodalite, which in turn is replaced by the unspecified Na-Al phase. b) Brownish carbonate aggregates, with the
inner parts being mainly composed of ankerite whereas the outer zones are dominated by carbocernaite. c) Latestage sulphide vein, composed of chalcopyrite, magnetite and millerite, with the latter being partially replaced by
fletcherite. d) Millerite in contact to chalcopyrite is replaced by a yet unknown member of the fletcherite group.
The REE are concentrated in conspicuous large, brownish carbonate aggregates, most
probably pseudomorphing the shape of a suspected subhedral precursor carbonate (presumably
calcite or ankerite), situated between large ankerite and magnetite grains. The carbonate
aggregates can be subdivided into at least two different zones. The inner, pale-coloured parts are
either composed of ankerite or consist of lamellar to patchy intergrowths of calcite and minor
194
13 Petrography
siderite. In clear contrast, the brownish outer zones of the domains are dominated by dark brown
carbocernaite (general formula: (Ca,Na)(Sr,Ce,Ba)(CO3)2; up to 15 vol.%) and ankerite (Fig.
3.10 b). It seems rather likely that these aggregates formed by metasomatic exchange between
ankerite and late-stage, REE-rich fluids. Late subhedral barite and anhedral strontianite grains of
up to 40 µm in diameter are localised in the interstices between large ankerite and magnetite
grains and the carbocernaite aggregates, testifying to autometasomatism of the rocks, caused by
the injection of late-stage Ba- and Sr-rich fluids.
Chalcopyrite and pyrite occur isolated in the matrix or, more commonly, as irregular
sulphide veins or patches, where the sulphides are associated with magnetite (Fig. 13.10 c).
Millerite is present in only minor proportions and is partially replaced by chalcopyrite.
Additionally, millerite is altered along its margins and along cracks to a yet unknown member of
the fletcherite group (Fig. 13.10 d). These fletcherites differ from those yet described by
containing variable but high amounts of Cu, whereas the Co and Fe contents are negligible.
13.5.4 Crystallization sequence of the carbonatitic breccia
In summary, the following magmatic and metasomatic crystallization sequence can be
established for the sodalite-rich carbonatitic breccia:
(1) In the cementing carbonatitic matrix magnetite, apatite, pyrochlore and biotite were
the first minerals to crystallize, followed by ankerite, the suspected precursor pyrrhotite and late
interstitial calcite. Chalcopyrite and pyrite are most probably synchronous with the second
magnetite generation. During the end-stages of, and outlasting the carbonatite crystallization and
fenitizing processes millerite, polydymite/violarite, chalcocite/digenite, bornite as well as late
hematite, chlorite, barite, strontianite and calcite of the second generation were formed.
(2) In layers composed of fragmented fenitized wallrock anorthosite, albite is the first
mineral to be formed during fenitization. Since albite is commonly included in apatite and biotite
of the cementing carbonate matrix, whereas both An-rich plagioclase and K-feldspar have never
been observed, it can be concluded that the transformation of feldspars of the wallrock
anorthosite and syenite into almost pure albite predated the emplacement of the carbonatitic
melt. The absence of Fe-Mg silicates common in anorthosite and syenite (e.g. pyroxene, olivine
and/or amphibole) provides evidence that biotite, like albite, was formed during an early
fenitizing event. After the emplacement of the carbonatitic melts, the repeated release of
fenitizing fluids caused the formation of sodalite at the expense of albite. This process started
13 Petrography
195
during, and outlasted the crystallization of biotite in the ankerite-rich zones. The replacement of
albite by muscovite and cancrinite started during the final stages and outlasted the formation of
sodalite in the fenitic zones and of the second magnetite generation in the carbonatite matrix.
The end-stages of the fenitization are marked by the formation of chlorite, hematite, rutile and
epidote.
13.6 Ferrocarbonatite veins
The younger ferrocarbonatite dykes and veins transsect both the massive and the banded variety
of the main carbonatite body and are almost free of wallrock-material, even though the veins
may locally be intermingled with the carbonatitic breccia. The following samples have been
investigated in detail:
Ferrocarbonatite:
Ku-97-15f, -15h, -15j, -15m, -15p, Ku-98-32, Ku-98-34, Ku-98-91, Ku-01-03,
Ku-01-04, Ku-01-05
Major phases of the massive ferrocarbonatites are subhedral to anhedral ankerite grains
with curved grain boundaries together with subordinate, subhedral to anhedral, angular to
irregularly shaped magnetite grains that frequently display albite inclusions (Fig. 13.11 a).
Martitisation of magnetite is mainly initiated from grain margins. Minor calcite fills the
interstices between ankerite grains whereas rutile occurs as rims surrounding magnetite (Fig.
13.11 b). In one sample optically zoned pyrochlore euhedra, containing inclusions of pure albite
(An0), have been observed in the vicinity of magnetite grains (Fig. 13.11 c). The contacts
between the ferrocarbonatite veins and the carbonatitic breccia are locally marked by trails of
euhedral magnetite grains within the ferrocarbonatite. In two samples, cancrinite rims, up to 2
mm across, are developed between the Si-poor ferrocarbonatite and the main, xenolith-rich
carbonatite intrusion, pointing to a minor degree of metasomatic exchange between the two rock
units. Rounded fragments of sodalite and mostly untwinned albite (An0-1) with undulose
extinction, derived from the fenitized wallrock anorthosite and/or fenite xenoliths within the
carbonatitic breccia, may form a subordinate constituents of this rock type (Fig. 13.11 d).
Thompson et al. (2002) additionally described the presence of quartz within the ferrocarbonatite
veins, incorporated from the fenitized wallrock anorthosite. In clear contrast to these authors,
quartz was never observed in the samples investigated in this study, comprising both the
196
13 Petrography
carbonatitic breccia and the younger ferrocarbonatite veins. The lack of quartz agrees well with
the general absence of quartz in the anorthosites of the KIC (Drüppel et al., 2001) and their
Fig. 13.11: Polished-section photomicrographs of the ferrocarbonatite veins (a-d) and the feldspar-biotite-ilmenitecalcite rock (e-f). Cross-polarized light (d). Plane-polarized light (a-c, e-f). a) Granular ankerite and subhedral to
euhedral magnetite, the latter bearing albite inclusions. b) Broad rutile rims are developed around magnetite. c)
Optically zoned, euhedral pyrochlore, bearing an albite inclusions. d) Fragmented relict of albite, displaying
undulose extinction. e) Homogeneous K-feldspar grains and anhedral to subhedral biotite flakes are the main
constituents of the feldspar-biotite-ilmenite-calcite rock. f) Biotite, bearing abundant inclusions of euhedral calcite.
13 Petrography
197
fenitization products and its rare occurrence in the syenites.
13.7 K-feldspar-biotite-ilmenite-calcite rock
The K-feldspar-biotite-ilmenite-calcite rock (Ku-98-25) forms part of a large dyke of the
sodalite-rich carbonatitic breccia. The black, medium-grained rock displays curved contacts with
the carbonatitic breccia and, in clear contrast to the latter, is characterized by a massive
appearance.
The main minerals present are medium-grained, subhedral to anhedral, tabular K-feldspar
(55 vol.%), mostly lacking exsolution phenomena, and subhedral to anhedral biotite flakes (35
vol.%; Fig. 13.11 e). Porphyric biotite was observed locally. Calcite (8 vol.%) forms either (1)
discrete, subhedral to anhedral grains in the matrix, filling K-feldspar interstices and displaying
straight grain boundaries against biotite or (2) euhedral inclusions in the latter (Fig. 13.11 f).
Anhedral ilmenite showing abundant lamellar twins after {10-11} is present in minor amounts,
filling the interstices between individual K-feldspar crystals. Accessory, fine-grained pyrite is
dispersed in the matrix.
The rock differs significantly from the investigated fenites in containing K-feldspar as a
major constituent and deviates from the syenites by the lack of exsolution features of potassium
feldspar and its comparably high modal amount of magmatic calcite. It can thus be concluded,
that the K-feldspar-biotite-ilmenite-calcite rock represents an independent magmatic unit,
displaying similarities with lamprophyres.
198
14 Mineral chemistry
The mineral chemistry of 27 selected samples of the rock types described in Chapter 13 has been
investigated with the aid of electron-microprobe (EMP) analysis. Analytical conditions are
provided in the Appendix A.2.3.
Fenitized anorthosite (A,f):
Ku-98-71, Ku-98-78, Ku-98-79, Ku-98-84, Ku-99-02
Fenitized syenite (S,f):
Ku-98-59b
Fenitized nepheline syenite (S,ne):
Ku-99-15a
Fenitized ultramafic xenolith (UMX):
Ku-97-43a
Sodalite-rich carbonatitic breccia (CB,so): Ku-97-15, Ku-97-15a, Ku-97-15d, Ku-97-15e, Ku-9719, Ku-97-24, Ku-98-57a, Ku-98-58a, Ku-98-80, Ku98-131, Ku-99-05, Ku-99-SA8
REE-rich samples (CB,REE):
Ku-98-130a, Ku-98-130b
Layered carbonatitic breccia (CB,l):
Ku-98-14, Ku-98-48, Ku-98-56b
Ferrocarbonatite veins (FV):
Ku-01-04
Kfs-Bt-Ilm-Cal rock (L):
Ku-98-25
14.1 Carbonate minerals
Carbonates are the main constituents of the ferrocarbonatite veins, the fenitized ultramafic
xenoliths and the cementing carbonatite matrix of the carbonatitic breccia. In addition, they are
ubiquitous accessory phases of the Kfs-Bt-Ilm-Cal rock and of both the anorthositic and fenitized
syenitic lithologies. The calculation of the ankerite, dolomite and carbocernaite formulae was
based on 2 cations whereas calcite, siderite, magnesite and strontianite were calculated on a 1cation basis. Representative analyses are listed in Tables A.5.1.13 to A.5.1.17 in the Appendix
A.5.1.
14.1.1 Ankerite-dolomite solid solution
The vast majority of the carbonates analysed has an Mg/(Mg+Fe) ratio of < 0.8 and can thus be
classified as ankerites. The MnO contents of the carbonates are generally high, whereas their Ce,
199
Ca
Ca
interstitial
calcite
interstitial
calcite
matrix
grains
euhedral
matrix grains
a
CB,so (N=75)
Ku-98-80
CB,so (N=46)
Ku-97-43a
Mg
b
(Fe+Mn)
Mg
(Fe+Mn)
Ca
Ca
matrix
grains
matrix
grains
CB,so (N=73)
Ku-98-131
CB,so (N=69)
Ku-99-05
c
d
Mg
(Fe+Mn)
Mg
(Fe+Mn)
Ca
Ca
Agg
matrix
grains
matrix
grains
CB,REE (N=76)
Ku-98-130a
Ku-98-130b
CB,so (N=143)
Ku-99-SA8
e
Mg
f
(Fe+Mn)
Mg
Agg
(Fe+Mn)
Fig. 14.1: Compositional variability of carbonate from the sodalite-rich carbonatitic breccia, expressed in the
ternary diagram Ca-Mg-(Fe+Mn). (Agg carbonates from brownish carbonate aggregates).
200
La and Na contents are near or below the detection limit. Representative analyses are listed in
Table A.5.1.13 in the Appendix A.5.1; average compositions of ankerite/dolomite of the
investigated samples are compiled in Table 14.1.
Carbonates with XMg at or near the ankerite-dolomite transition were observed in two
samples of the carbonatitic breccia, where they occur in different textural positions: (1) as
euhedral, early crystallized carbonates (XMg: 0.61-0.77) in the sodalite-rich carbonatitic breccia,
enclosing abundant ultramafic xenoliths (Ku-97-43a; Fig. 14.1a) and (2) as carbonate inclusions
(XMg: 0.79-0.84) in magnetite of the first generation in sample Ku-98-14 of the sodalite-free
carbonatitic breccia (Fig. 14.2a). In clear contrast, anhedral matrix carbonates of the main
carbonatite body display distinctly lower Mg-contents, that slightly decrease in the sequence of
layered carbonatitic breccia (XMg: 0.56-0.74; Fig. 14.2) – REE-poor samples of the sodalite-rich
carbonatitic breccia (XMg: 0.43-0.64; Fig. 14.1a-e) - REE-rich samples of the sodalite-rich
carbonatitic breccia (XMg: 0.25-0.61; Fig. 14.1f). All analysed ankerite crystals show minor
oscillatory to patchy zoning with the Mg-contents commonly decreasing from core towards the
rim (Fig. 14.4a-d). Only the biotite-free sample Ku-99-05 (Fig. 14.4e) displays an opposite
zoning, with the Fe-content decreasing towards the rim. The observed trends of decreasing XMg
of the carbonates with subsequent crystallization most probably reflects the extensive,
synchronous biotite formation in these samples. In the same direction both MnO and SrO
increase from values 0.6 to 2.5 wt.% and 0.1 to 0.8 wt.%, respectively.
Ca
Ca
matrix
overgrowths
matrix
grains
matrix
grains
inclusion
in Mag
CB,l (N=48)
Ku-98-56b
CB,l (N=191)
Ku-98-14
a
Mg
b
(Fe+Mn)
Mg
(Fe+Mn)
Fig. 14.2: Compositional variability of carbonate from the layered carbonatitic breccia, expressed in the ternary
diagram Ca-Mg-(Fe+Mn).
In clear contrast, the rimward Fe-increase of the biotite-free sample Ku-99-04 is
suggested to reflect the onset of magnetite crystallization.
201
Ankerite from samples of the younger ferrocarbonatite veins (XMg: 0.43-0.64; Fig. 14.3a)
has Mg amounts comparable to those of ankerites from the sodalite-rich carbonatitic breccia, but,
at the same time, distinctly higher MnO contents of 2.3-3.4 wt.%, suggesting its crystallization
from a more fractionated melt. In clear contrast to carbonates of the main carbonatite body, most
of these ankerites preserve nearly unzoned Fe-rich cores (XMg: 0.43-0.49) and display strongly
reversed rims (XMg: 0.55-0.64) where the Mg content increases abruptly (Fig. 14.4f). This
increase is suggested to be related to the onset of magnetite crystallization in the
ferrocarbonatite, which in contrast to the carbonatitic breccia, lacks an Mg-fractionating phase
like biotite.
Ca
Ca
matrix
grains
matrix
grains
rims
FV (N=106)
Ku-01-04
A,f (N=186)
Ku-99-02
a
b
Mg
(Fe+Mn)
Mg
(Fe+Mn)
Ca
Ca
cracks
Bt-Ank
patches and coronas
inclusion
in Pl
UMX (N=107)
Ku-97-43a
S,f (N=58)
Ku-98-59b
c
Mg
d
(Fe+Mn)
matrix
grains
Mg
(Fe+Mn)
Fig. 14.3: Compositional variability of carbonate, expressed in the ternary diagram Ca-Mg-(Fe+Mn). a)
Ferrocarbonatite vein. b) Fenitized anorthosite. c) Fenitized syenite. d) Fenitized ultramafic xenolith.
Carbonates in the bordering fenitized anorthosites (Ku-99-02) are ankerites with a patchy
zoning and Mg/Fe ratios (XMg: 0.48-0.64) in the range of the main carbonatite body (Fig. 14.3b),
202
1.0
1.0
CB,l
Ku-98-14
0.9
0.8
0.7
0.7
0.6
0.6
(Fe+Mn)
(Fe+Mn)
0.8
0.5
0.4
0.5
0.4
0.3
0.3
0.2
0.2
a
0.1
CB,so
Ku-97-43a
0.9
b
0.1
0.0
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0
0.1
0.2
0.3
0.4
Mg
1.0
CB,so
Ku-98-80
0.8
0.7
0.8
0.9
1.0
0.8
0.7
0.7
0.6
0.6
0.5
0.4
0.5
0.4
0.3
0.3
0.2
0.2
c
0.1
CB,REE
Ku-98-130a
0.9
(Fe+Mn)
(Fe+Mn)
0.6
1.0
0.9
d
0.1
0.0
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0
0.1
0.2
0.3
0.4
Mg
0.5
0.6
0.7
0.8
0.9
1.0
Mg
1.0
1.0
CB,so
Ku-99-05
0.9
0.8
0.8
0.7
0.7
0.6
0.6
0.5
0.4
0.5
0.4
0.3
0.3
0.2
0.2
e
0.1
FV
Ku-01-04
0.9
(Fe+Mn)
(Fe+Mn)
0.5
Mg
f
0.1
0.0
0.0
0.0
0.1
0.2
0.3
0.4
0.5
Mg
0.6
0.7
0.8
0.9
1.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Mg
Fig. 14.4: Core (white symbols) and rim (black symbols) compositions of ankerite from selected samples of the
carbonatitic breccia (a-e) and the younger ferrocarbonatite veins (f). Note that biotite-rich samples (a-d) display a
rimward increase of the Fe-content, whereas biotite-free samples (e-f) are characterized by an opposite trend of
increasing Mg.
indicating that the wallrock-fenitization is related to the injection of carbonatite melt. With a
Mg/(Mg+Fe) of 0.27-0.34, narrow ankerite rims on the matrix ankerite grains display the highest
sample; wt.%
203
textural position
MgO
CaO
MnO
FeO
SrO
BaO
CO2
Sum
Ku-98-14
σ (N=4)
inclusion in Mag
15.80
27.99
1.96
6.68
0.11
0.02
44.59
97.16
0.38
0.17
0.64
0.80
0.04
0.02
0.42
0.83
Ku-98-14
σ (N=187)
anhedral matrix grains
11.30
27.00
2.40
12.60
0.80
0.03
43.07
97.19
0.57
0.34
0.16
0.89
0.10
0.04
0.36
0.72
Ku-98-56b
σ (N=4)
13.62
29.21
1.90
10.12
0.70
0.00
45.47
101.03
0.20
0.41
0.13
0.21
0.11
0.00
0.21
0.41
14.97
28.82
0.57
9.11
0.22
0.03
44.99
98.70
0.77
0.43
0.09
1.24
0.07
0.03
0.39
0.76
28.04
1.60
13.74
0.77
0.03
44.44
100.25
CB,l
CB,so
Ku-97-43a
σ (N=31)
euhedral matrix grains
Ku-98-80
σ (N=72)
11.62
0.53
0.26
0.08
0.74
0.08
0.04
0.30
0.62
Ku-98-131
σ (N=73)
11.63
28.60
1.25
13.76
0.59
0.03
44.60
100.46
0.46
0.35
0.19
0.58
0.12
0.04
0.34
0.78
Ku-99-SA8
σ (N=143)
11.52
26.79
2.49
13.72
0.71
0.03
43.87
99.13
0.23
0.35
0.16
0.34
0.08
0.04
0.27
0.63
Ku-99-05
σ (N=69)
8.51
27.00
2.30
16.52
0.64
0.03
42.31
97.30
0.96
0.50
0.28
1.51
0.12
0.04
0.61
1.10
100.24
CB,REE
Ku-98-130a
σ (N=53)
Ku-98-130b
σ (N=28)
9.41
28.27
1.75
16.41
0.53
0.02
43.84
1.81
0.47
0.17
2.89
0.17
0.03
0.66
0.91
10.31
28.58
1.70
15.03
0.60
0.04
41.64
98.24
0.65
0.55
0.16
0.65
0.35
0.05
0.49
0.60
FV
Ku-01-04
σ (N=106)
9.78
28.37
2.73
15.05
0.62
0.03
41.45
98.14
1.29
0.58
0.27
2.26
0.09
0.04
0.64
1.11
10.43
26.90
1.98
15.12
0.56
0.02
43.24
98.26
0.85
0.34
0.35
1.14
0.14
0.03
0.38
0.65
5.64
26.39
1.30
23.17
0.45
0.04
42.07
99.06
0.72
0.26
0.00
0.69
0.06
0.04
0.53
0.72
A,f
Ku-99-02
σ (N=184)
Ku-99-02
rims on matrix grains
σ (N=2)
S,f
Ku-98-59b
σ (N=11)
6.14
29.85
2.14
16.66
0.46
0.02
41.86
97.11
1.02
3.59
0.38
2.64
0.12
0.02
0.65
1.23
15.80
28.27
0.43
8.77
0.37
0.03
45.24
98.91
0.78
0.69
0.09
1.02
0.08
0.04
0.35
0.65
UMX
Ku-97-43a
σ (N=68)
patches and coronas
Table 14.1: Average compositions of ankerite/dolomite from different samples and textural positions (σ standard
deviation).
Fe-contents of all ankerites analysed, and, moreover, seem to extend the yet known
compositional range of natural ankerite to the most extreme Fe-rich composition (Fig. 14.3b).
204
Ankerite of the fenitized syenite (Ku-98-59b) has a comparably low XMg (0.36-0.44; Fig.
14.3c), according to high MnO amounts of 1.9-2.9 wt.%. The Fe-rich nature of the carbonates
reflects the Fe-rich whole-rock composition of the syenites, and thus suggests some sort of
chemical interaction between the intruding carbonatite melt and the bordering lithologies during
fenitization.
Ankerite/dolomite from the fenitized ultramafic xenolith (Ku-97-43a), occurring in
biotite-ankerite patches, schlieren and corona textures, has quite homogeneous MgO contents
(XMg: 0.71-0.81; Fig. 14.3d), which, however, are higher than those of ankerite/dolomite in the
enclosing carbonatitic breccia (Fig. 14.1a). Like in the case of the fenitized syenite, these
changes in the mineral compositions of ankerite/dolomite are suggested to reflect chemical
interaction between the carbonatite melt and the high-Mg ultramafic xenoliths.
14.1.2 Calcite
Calcite of the carbonatitic breccia occurs in 3 different textural positions, i.e. (1) in the
interstices between individual ankerite grains (Ku-97-43a, Ku-98-14, Ku-98-80, Ku-98-130a),
2.6
matrix
grains
2.4
2.2
2.0
SrO (wt.%)
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0
S, f (Ku-98-59)
L (Ku-98-25)
inclusion
in Bt
te
lci
a
cc
i
t
a
ite
c
l
m
ca
ag
d
m
e
z
y
ar
ili
b
m
carbonate
o
ri
m
aggregates p
re
y
all
rm
he
t
o
dr
y
h
matrix
overgrowths
cracks,
rims around Mag
matrix
grains
matrix
grains
0.2
CB, so (Ku-98-80)
CB, so (Ku-97-43a)
CB, REE (Ku-98-130a)
CB, l (Ku-98-14)
CB, l (Ku-98-56)
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
MnO (wt.%)
Fig. 14.5: Compositional variability of calcite, plotted as SrO against MnO (CB,l layered carbonatitic breccia,
CB,REE REE-rich samples of the sodalite-rich carbonatitic breccia, CB,so sodalite-rich carbonatitic breccia, L Kfeldspar-biotite-ilmenite-calcite rock, S,f fenitized syenite).
205
(2) as late matrix overgrowths (Ku-98-56b) and (3) in brownish carbonate aggregates, where it is
intimately associated by siderite and carbocernaite (Ku-98-130a). Orthomagmatic, interstitial
calcites of samples Ku-98-14 (MnO: <0.08, SrO: 1.9-2.1), Ku-98-80 (MnO: <0.02, SrO: 1.6-2.2)
and Ku-98-43a (MnO: 0.1-0.3, SrO: 0.4-0.7) are characterized by high Sr/Mn ratios (Fig. 14.5).
Presumably also orthomagmatic is calcite from the brownish carbonate aggregates (MnO: 0.20.3, SrO: 0.7-0.8) of the REE-rich carbonatitic breccia Ku-98-130a (Fig. 14.5).
sample; wt.%
textural position
MgO
CaO
MnO
FeO
SrO
BaO
CO2
Sum
cracks, rims around
0.56
51.27
1.48
2.42
0.57
0.06
43.52
99.89
Mag, matrix
0.31
1.09
0.16
0.93
0.06
0.06
0.34
0.81
100.26
S, f
Ku-98-59b
σ (N=48)
Kfs-Bt-Ilm-Cal rock
Ku-98-25
σ (N=8)
inclusions in Bt
Ku-98-25
σ (N=10)
0.01
54.52
0.04
0.16
1.80
0.03
43.70
0.02
0.32
0.04
0.10
0.13
0.04
0.24
0.58
0.99
50.76
0.76
1.95
2.09
0.05
43.49
100.09
0.30
0.22
0.06
0.12
0.10
0.06
0.34
0.67
99.68
CB,so
Ku-97-43a
σ (N=15)
Ku-98-80
σ (N=3)
0.03
54.31
0.19
0.99
0.51
0.04
43.61
0.04
0.48
0.07
0.11
0.08
0.04
0.43
0.99
0.00
53.43
0.01
0.08
1.66
0.00
42.69
97.86
0.00
0.11
0.01
0.06
0.25
0.00
0.09
0.31
99.42
CB,REE
Ku-98-130a
σ (N=2)
brownish carbonate
0.12
54.31
0.23
0.37
0.76
0.14
43.49
aggregates
0.03
0.05
0.06
0.02
0.03
0.03
0.05
0.14
Ku-98-130a
0.20
51.69
0.91
2.95
0.67
0.07
43.46
99.95
0.55
4.71
0.31
4.13
0.60
0.06
0.57
0.98
98.82
σ (N=21)
CB,l
Ku-98-14
σ (N=4)
Ku-98-56b
matrix overgrowths
σ (N=44)
0.00
52.86
0.05
0.10
2.00
0.06
41.03
0.00
0.54
0.03
0.07
0.04
0.07
0.37
0.97
0.08
53.23
1.28
0.79
0.96
0.03
43.56
99.93
0.08
0.61
0.13
0.27
0.19
0.04
0.34
0.80
Table 14.2: Average compositions of calcite from different samples and textural positions (σ standard deviation).
In clear contrast, anhedral matrix calcite of the REE-rich carbonatitic breccia Ku-98130a, has comparably high MnO contents of 0.7-1.0 wt.% and intermediate SrO contents of 0.30.9 wt.%, well in accordance with a secondary origin of calcite during hydrothermal alteration
(Fig. 14.5). Similar Sr/Mn ratios have been obtained for texturally late calcite of sample Ku-9856b (MnO: 1.0-1.6 wt.%, SrO: 0.5-1.3 wt.%).
Calcite is the dominant mineral among carbonates in the fenitized syenite, where the
carbonate occurs in cracks within feldspar and as rims around magnetite. With MnO contents of
206
1.2-1.8 wt.% and SrO contents of 0.5-0.6 wt.%, it has the lowest Sr/Mn of all calcites analysed,
and is thus presumably of secondary origin (Fig. 14.5).
Calcite of the K-feldspar-biotite-ilmenite-calcite rock has been observed in two textural
positions, (1) as inclusions in biotite and (2) as anhedral matrix grains. Calcite of both textures
has constantly high SrO contents (1.6-2.2 wt.%) but quite variable MnO values, ranging between
0.0–0.1 wt.% in calcite inclusions and 0.7–0.9 wt.% in matrix grains. However, calcite of both
generations is presumably of magmatic origin (Fig. 14.5).
14.1.3 Siderite-magnesite solid solution
Siderite was observed in the central parts of the brownish carbonate aggregates of the REE-rich
carbonatitic breccia. Since siderite has grown close to calcite, only one reliable analysis was
obtained, whereas all others revealed mixtures between calcite and siderite. Siderite is
characterized by a high XFe of 0.81 and high MnO and CaO contents of 3.3 wt.% and 1.35 wt.%,
respectively (Fig. 14.6), whereas Sr and Ba are below the detection limit. Siderite is suggested to
be exsolved from a Fe-rich precursor carbonate (presumably calcite or ankerite) during cooling.
Mg (+ Ca)
UMX (N=39)
Ku-97-43a
CB,REE (N=1)
Ku-98-130a
Fe
Mn
Fig. 14.6: Compositional variability of magnesite and siderite, expressed in the ternary diagram Mg(+Ca)-Fe-Mn.
Note that both magnesite and siderite contain only minor Ca with respect to Mg.
Magnesite is a major constituent of the fenitized ultramafic xenoliths, where it is
intimately associated with aegirine and magnesio-arfvedsonite. All magnesites analysed contain
between 20-38 mol.% FeCO3 and can thus be classified as breunnerite, the ferroan magnesite
207
variety (Fig. 14.6). The breunnerites contain only minor amounts of Sr and Ba, whereas MnO
and CaO range between 0.7-1.1 % and 0.2-0.3 wt.% , respectively. All analysed breunnerites
sample; wt.%
textural position
MgO
CaO
MnO
FeO
SrO
BaO
CO2
Sum
6.29
1.35
3.34
47.84
0.00
0.02
39.31
98.15
CB,REE
Ku-98-130a
carbonate aggregates
UMX
Ku-97-43a
σ (N=39)
26.08
0.25
0.88
27.17
0.02
0.03
45.88
100.31
1.12
0.04
0.09
1.49
0.03
0.03
0.44
0.77
Table 14.3: Average compositions of members of the siderite-magnesite solid solution series (σ standard deviation).
exhibit a weak concentric zonation, with the MgO contents slightly decreasing from core
towards the rim (XMg: 0.60-0.63→ XMg: 0.64-0.68). The Mg-rich composition of the analysed
magnesites strongly suggests their formation by alteration of former Fe-Mg silicates (presumably
olivine or its alteration products), following the injection of CO2-rich fluids of the carbonatite
melt. Judging from the zonation patterns of breunnerites, these fluids must have contained
certain amounts of Fe. The obtained results strongly suggest, that the fenitization of the
ultramafic xenoliths was caused by the Fe-rich carbonatitic melts, which transported them into
upper crustal levels.
14.1.3 Carbocernaite
Carbocernaite (Ca,Na)(Sr,Ce,La)(CO3)2 has been analysed in two REE-rich samples of the
carbonatitic breccia (Ku-98-130a, Ku-98-130b). It has been distinguished from the chemically
similar mineral burbankite (Na,Ca,Sr,Ba,Ce)6(CO3)5 by X-ray diffraction and its stoichiometry,
which is rather 6:6 than 6:5 regarding the cation:carbonate ratio.
The analysed carbocernaite grains (Table A.5.1.16 in the Appendix A.5.1) have a rather
variable composition with the CaO and SrO contents ranging between 14.3-18.1 wt.% and 13.618.6 wt.%, respectively. In most cases the REE carbonates display oscillatory zoning patterns.
Decreases of Ca and Sr in the analyses are generally accompanied by increases of the Na2O (1.22.2 wt.%), Ce2O3 (15.1-20.1 wt.%) and La2O3 (10.7-14.4 wt.%) contents. The compositions of
the
analysed
carbocernaite
range
from
(Ca0.97Na0.10Sr0.50Ba0.03Ce0.23La0.17)(CO3)2.
(Ca1.01Na0.09Sr0.58Ba0.06Ce0.15La0.11)(CO3)2
and
208
The chemistry of the analysed carbocernaites is in good agreement with that obtained by
Wall et al. (1993) for carbocernaite exsolution lamellae in calcite in a REE-rich carbonatite dyke
from Rajasthan, India. The close association of carbocernaite with calcite-siderite intergrowths
might suggest that carbocernaite, like siderite, was exsolved from a high-REE precursor calcite,
which additionally contained a certain amount of Fe. However, a cotectic crystallization of
calcite and carbocernaite, followed by an exsolution of siderite-component of calcite can not yet
be ruled out.
14.1.4 Strontianite
Strontianite is a rare mineral in the Swartbooisdrif carbonatitic breccia. It was observed as rims
around carbocernaite in the marginal parts of the brownish carbonate aggregates of the REE-rich
sample Ku-98-130a. The only reliable analysis of the very fine-grained strontianite (Table
A.5.1.17 in the Appendix A.5.1) revealed exceptionally high CaO contents of 18.3 wt. %, which
is equivalent to 42 mol.% CaCO3. The relation of strontianite to the carbocernaite remains
somewhat unclear, however, detailed EMP analysis and backscattered imagery on the
conspicuous fine-grained carbonate aggregates is in progress.
14.2 Feldspar
Representative analyses of potassium feldspar and plagioclase are depicted in Table A.5.1.1 in
the Appendix A.5.1.
14.2.1 Potassium feldspar
K-feldspar is a major orthomagmatic phase of the nepheline syenite (Ku-99-15a), where it forms
large, subhedral grains. The feldspar grains exhibit strong perthitic exsolution, with a potassium
feldspar host (Or96-97), bearing plagioclase (Ab98-100) exsolution lamellae (Fig. 14.7a). The Anand Ce-contents of both the K-feldspar host and the albite exsolution lamellae are generally
below 0.5 mol.%. Due to the coarse grain sizes of the exsolution lamellae, reintegrating
measurements with a broad, defocused beam yielded no reliable results.
209
Unexsolved potassium feldspar is the main constituent of the K-feldspar-biotite-ilmenitecalcite rock. The K-feldspar is characterized by a weak normal zonation, with both the Ab and
the Ce contents increasing from core (Or96.3Ab3.7Ce0.0-Or92.6Ab4.9Ce2.5) towards the rim
(Or91.1Ab5.9Ce3.0-Or86.0Ab8.2Ce5.8), whereas the An contents are generally below the detection
limit (Fig. 14.7b).
An
An
focused beam
(beam size: 1 µm)
Ku-97-15a
focused beam
(beam size: 1 µm)
Ku-98-25
a
b
Or
Ab
Ab
Or
Fig. 14.7: Compositional variability of potassium feldspar of (a) the nepheline syenite and (b) the K-feldsparbiotite-ilmenite-calcite rock, expressed in the ternary diagram An-Ab-Or.
Feldspar relics of former microcline, observed in the fenitized syenite sample Ku-98-59b,
are virtually pure albite. Despite the fact, that most feldspar fragments were extensively
analysed, not one single K-feldspar analysis was obtained, suggesting that the fenitization of the
syenite resulted in a complete replacement of K-feldspar by chess-board albite.
14.2.2 Plagioclase
Metasomatically formed feldspar from the fenitized syenite, the carbonatitic breccia and the late
ferrocarbonatite veins is almost pure, homogeneous albite (Ab99-100; Table 14.4). The same
compositional range was obtained for albite inclusions in pyrochlore of the ferrocarbonatite
veins (Table 14.4). The homogeneity of the values is striking, hence suggesting that all above
mentioned lithologies must have re-equilibrated with large amounts of Na-rich fluids, expelled
by the carbonatite magma.
210
sample
SiO2
Al2O3
CaO
FeO
BaO
Na2O
K2O
Sum
An
Ab
Ku-97-15
σ (N=33)
68.77
20.08
0.02
0.61
0.32
0.03
0.07
0.02
0.06
0.02
Ku-98-57a
σ (N=38)
67.76
19.55
0.05
0.83
0.24
0.05
0.21
0.02
0.83
0.03
Ku-98-80
σ (N=28)
69.02
0.45
20.25
0.31
0.03
0.03
0.29
0.33
Ku-98-131
σ (N=14)
69.18
20.21
0.03
0.79
0.30
0.02
Ku-99-05
σ (N=21)
67.79
19.41
0.04
0.25
0.11
0.03
68.26
20.15
0.02
0.64
0.07
0.02
Ku-98-14
σ (N=24)
67.28
0.35
19.35
0.14
0.03
0.06
0.11
0.09
Ku-98-48
σ (N=18)
68.95
20.27
0.03
0.60
0.35
0.03
Ku-98-56
68.78
20.22
67.64
0.61
68.05
20.23
0.06
0.78
0.39
0.07
Ku-99-02
σ (N=13)
68.63
20.25
0.06
0.59
0.29
Ku-98-71
68.29
20.33
Ku-98-71
Ku-98-79
66.68
63.17
19.95
22.17
Ku-98-79
63.73
Ku-98-79
66.79
Ku-98-79
68.22
Or
Ce
11.66
0.07
100.68
0.12
99.48
0.37
0.03
0.15
0.03
0.77
0.14
0.19
0.16
0.04
11.65
0.04
99.31
0.23
99.50
0.24
0.03
0.15
0.02
0.50
0.21
0.21
0.10
0.05
0.03
0.03
11.50
0.11
0.07
0.02
101.20
0.57
0.16
0.16
99.38
0.22
0.41
0.13
0.05
0.05
0.08
0.03
11.45
0.07
101.13
0.13
99.41
0.41
0.05
0.17
0.02
0.24
0.02
0.73
0.09
0.12
0.09
0.03
0.11
0.03
11.57
0.06
99.02
0.17
99.45
0.32
0.05
0.04
0.03
0.12
0.02
0.33
0.15
0.22
0.14
0.05
0.04
0.01
11.56
0.05
100.10
0.10
99.60
0.30
0.01
0.01
0.01
0.17
0.01
0.70
0.08
0.15
0.05
0.01
0.02
0.03
11.60
0.17
0.06
0.02
98.47
0.47
0.16
0.29
99.46
0.29
0.34
0.13
0.04
0.05
0.18
0.02
11.52
0.09
101.07
0.14
99.30
0.52
0.04
0.18
0.02
0.26
0.02
0.88
0.13
0.21
0.14
0.04
0.01
0.02
0.00
11.73
0.05
100.82
0.03
99.68
0.29
0.00
19.45
0.04
0.18
0.02
11.56
0.05
98.95
0.18
99.49
0.30
0.00
0.25
0.02
0.18
0.02
0.19
0.02
0.54
0.11
0.09
0.10
0.00
0.05
0.03
11.50
0.06
99.99
0.30
99.32
0.33
0.05
0.08
0.03
0.31
0.04
0.87
0.36
0.41
0.21
0.06
0.03
0.06
11.39
0.05
100.49
0.28
99.34
0.27
0.10
0.05
0.02
0.04
0.17
0.03
0.74
0.24
0.27
0.15
0.07
0.32
0.09
0.14
11.46
0.03
100.68
1.52
98.06
0.18
0.25
1.92
3.04
0.14
1.09
0.00
0.08
11.87
8.69
0.01
0.02
100.56
98.30
8.19
16.13
91.77
83.59
0.05
0.13
0.00
0.15
22.32
3.37
0.10
0.00
8.84
0.10
98.49
17.29
82.08
0.63
0.00
20.88
1.33
0.03
0.04
9.92
0.03
99.01
6.88
92.87
0.18
0.07
20.40
0.74
0.05
0.04
10.23
0.05
99.72
3.82
95.82
0.29
0.07
CB,so
CB,REE
Ku-98-130a
σ (N=2)
CB,l
FV
Ku-01-04
σ (N=15)
S, f
Ku-98-59b
σ (N=139)
A, f
Table 14.4: Average compositions of plagioclase (σ standard deviation).
Magmatic plagioclase (An37-75) of the fenitized anorthosites (Ku-98-71, Ku-98-79)
bordering the carbonatitic breccia is replaced, to an increasing extent, by sodium-rich plagioclase
(An1-17). At the direct contact to the carbonatite dykes, fenitization reached pervasive extent. In
the anorthositic fenite sample Ku-99-02, taken from a foliated contact zone, the magmatic
plagioclase is completely transformed into homogeneous and virtually pure albite (Ab99-100;
Table 14.4).
211
During the transformation of both, potassium feldspar of the syenite and of plagioclase of
the anorthosite into almost pure albite, appreciable amounts of K and Ca must have been
released, which most probably account for the close association of albite and muscovite in the
fenitized syenite and of albite and carbonate in the fenitized anorthosite.
14.3 Nepheline
Euhedral to subhedral nepheline is a major constituent of the nepheline syenite, whereas
anhedral, metasomatically formed nepheline is an accessory phase. Formulae of nepheline were
calculated on a basis of 32 oxygens (Table A.5.1.18 in the Appendix A.5.1).
The analysed compositions of magmatic nepheline are in the range of Ne68.4Ks19.1Qtz12.5
to Ne73.6Ks18.6Qtz7.8. In most cases the grains display a continuous growth zonation, with SiO2
decreasing towards the rim (Fig. 14.8). When approaching the outermost margins or cracks
nepheline is significantly richer in the Ne component (Ne73.1Ks21.4Qtz5.5 – Ne80.3Ks11.2Qtz8.5).
SiO2
80
nepheline syenite
Ku-99-15
core
rim
outermost margin, cracks
core
90
rim
100
NaAlSiO4
90
80
70
60
KAlSiO4
Fig. 14.8: Compositional variability of nepheline from the nepheline syenite, expressed in the ternary diagram SiO2NaAlSiO4-KAlSiO4.
These chemical variations suggest an interaction between the nepheline-syenite and the
Na-rich and Si-deficient carbonatite magma. Since both, the petrographical observations and the
analyses of the nepheline mineral chemistry point to continuous nepheline growth, it might be
possible that Na was introduced to the system before nepheline was fully crystallized, hence
pointing to a contemporaneous carbonatite-nepheline syenite emplacement.
212
Remarkably, secondary, metasomatically formed nepheline differs significantly from
those of magmatic origin in containing significant amounts of Ca (1.0-1.5 a.p.f.u.) in addition to
Na (6.5-6.9 a.p.f.u.), but only minor amounts of K (0.1-0.2 a.p.f.u). Presumably, secondary
nepheline was formed instead of cancrinite by the interaction between magmatic nephelines and
the Na-and Ca-rich carbonatitic magma.
14.4 Sodalite
In the Swartbooisdrif area, sodalite occurs in different rock types and textural positions,
i.e. (1) in the fenitized wallrock anorthosite, replacing plagioclase, (2) as a replacement product
of nepheline in the nepheline syenites, (3) as replacement product of feldspar-rich anorthositic
wallrock fragments, incorporated by the carbonatitic breccia, (4) in the late ferrocarbonatite
veins, replacing albite-xenocrysts and, quite less common, (5) in the fenitized syenite, replacing
chess-board albite. The sodalites from the different rock types have been analysed with the help
of EMP analysis. In order to prevent migration of sodium during electron bombardment, sodalite
has been measured with a broad beam of 10 µm across. Sodalite formulae were calculated on the
basis of 24 oxygens, assuming charge balance. Representative analyses are listed in Table
A.5.1.19 in the Appendix A.5.1, average analyses are provided in Table 14.5.
45
a
c
b
e
d
Cl, Na2O, Al2O3, SiO2 (wt.%)
40
35
30
25
20
SiO2
Al 2O3
Na2O
Cl
15
10
5
S, ne
CB, so (Ku-97-15)
CB, so (Ku-99-SA8)
CB, so
(98-130a)
CB, so
(Ku-98-131)
CB, so
(Ku-98-80)
Fig. 14.9: Zoning patterns of sodalite from the nepheline syenite (a) and the sodalite-rich carbonatitic breccia (b-e).
213
Sodalite of all rock units is the almost pure sodalite endmember, generally displaying
Fe2O3, CaO, K2O, S and F contents <0.1 wt.%. However, most of the analysed sodalites deviate
from the ideal sodalite composition Na8[Cl2(AlSiO4)6] in containing too low Na and too high Si
and
Al,
with
the
sodalite
formulae
ranging
between
Na7.0[Cl2.0(Al6.2Si6.1O24)]and
Na7.3[Cl2.0(Al6.2Si6.0O24)] in the nepheline syenite and between Na6.3[Cl2.0(Al6.4Si6.2O24)] and
Na7.7[Cl2.0(Al6.1Si6.0O24)] in the carbonatitic breccia. Remarkably, both Si and Al apparently
increase with decreasing Na, whereas Cl remains constant (Fig. 14.9).
7.8
8.0 Sdl
7.8
7.6
7.6
7.4
7.4
Sdl
Na (a.p.f.u.)
Na (a.p.f.u.)
8.0
7.2
7.0
6.8
7.2
7.0
6.8
6.6
6.6
6.4
6.4
6.2
6.2
a
6.0
5.90 5.95 6.00 6.05 6.10 6.15 6.20 6.25 6.30
b
6.0
6.00 6.05 6.10 6.15 6.20 6.25 6.30 6.35 6.40 6.45
Si (a.p.f.u.)
Al (a.p.f.u.)
6.45
6.40
fenitized Ne-syenite
6.35
Al (a.p.f.u.)
CB,so
Ku-97-15
Ku-98-57a
Ku-98-58a
Ku-98-80
Ku-98-130a
Ku-98-131
Ku-99-05
Ku-99-SA8
6.30
6.25
6.20
6.15
6.10
6.05
c
6.00
5.90
Sdl
5.95
6.00
6.05
6.10
6.15
6.20
6.25
Si (a.p.f.u.)
Fig. 14.10: Correlation patterns of Na, Si and Al p.f.u. (formula calculation based on 24 oxygens) in sodalite. a) Si
against Na. b) Al against Na. c) Si against Al. Sdl marks ideal sodalite composition.
Such a crystal chemical behaviour has so far never been observed for natural sodalites. The
apparent negative correlation of Na with both Al and Si (Fig. 14.10) may thus be related to the
conspicuous crystallization history of sodalite, i.e. its subsolidus formation at the expense of
albite under Si-deficient conditions. It may be possible, that albite was not completely
transformed into sodalite during fenitization, depending on the fluid:rock ratio and the local
214
availability of Na, Si and Al. This possibility, however, has to be examined with the help of
detailed TEM studies.
The transformation of albite of the fenitized anorthosites and syenites and of nepheline of
the nepheline syenites into sodalite necessitates the introduction of large amounts of NaCl,
whereas Si and K will be released, as is evidenced by the following two reactions (balanced for
Al):
5.91 Na0.98K0.02Si2.97Al1.03O8 + 1.86 Na+ + 2.00 Cl- → Na7.65Cl2.00Si6.02Al6.09O24 + 11.53 SiO2 + 0.12 K+
albite
in solution
sodalite
(14-1)
in solution
Na6.37K1.63Si8.01Al7.99O32 + 2.55 Cl- + 2.99 Na+ → 1.28 Na7.30Cl1.99Si5.99Al6.23O24 + 0.33 SiO2 + 1.63 K+
nepheline
in solution
sodalite
(14-2)
in solution
The release of K and SiO2 during the sodalitization of nepheline may account for the
close spatial association of sodalite and muscovite in the nepheline syenite, whereas it remains
sample
SiO2
Al2O3
Fe2O3
CaO
Na2O
K2O
Cl
F
S
Total
O-Cl,S,F
Sum
CB,so
Ku-97-15
σ (N=28)
38.03
0.24
32.92
0.35
0.03
0.03
0.02
0.02
24.31
0.22
0.02
0.02
7.30
0.14
0.00
0.00
0.02
0.03
102.65
0.53
1.68
0.05
100.97
0.52
Ku-97-43a
σ (N=4)
39.05
0.98
32.68
0.13
0.02
0.03
0.04
0.02
22.36
0.29
0.03
0.01
7.15
0.13
0.05
0.08
0.01
0.01
101.39
0.65
1.65
0.05
99.74
0.69
Ku-98-130a
σ (N=8)
39.21
0.38
33.95
0.58
0.03
0.02
0.02
0.02
21.13
0.67
0.03
0.02
7.47
0.07
0.04
0.05
0.01
0.01
101.89
0.56
1.72
0.03
100.17
0.54
Ku-98-131
σ (N=19)
38.92
0.40
33.64
0.33
0.05
0.17
0.02
0.02
21.17
0.89
0.01
0.03
7.42
0.10
0.04
0.05
0.02
0.02
101.29
0.53
1.72
0.04
99.57
0.53
Ku-98-57a
σ (N=11)
38.17
0.49
32.68
0.54
0.04
0.03
0.02
0.02
23.09
0.58
0.01
0.01
7.42
0.14
0.06
0.07
0.01
0.01
101.50
0.50
1.72
0.05
99.78
0.48
Ku-98-58a
σ (N=6)
38.16
0.19
33.42
0.24
0.31
0.21
0.02
0.02
22.69
0.16
0.04
0.02
7.47
0.07
0.05
0.05
0.01
0.01
102.16
0.45
1.72
0.02
100.44
0.46
Ku-98-80
σ (N=21)
38.58
0.59
33.44
0.69
0.01
0.02
0.01
0.02
21.87
1.11
0.02
0.01
7.55
0.11
0.04
0.05
0.02
0.01
101.55
0.46
1.74
0.04
99.81
0.47
Ku-99-05
σ (N=7)
38.03
0.89
32.23
0.51
0.02
0.02
0.01
0.01
22.51
0.73
0.06
0.01
7.15
0.20
0.09
0.06
0.01
0.01
100.12
0.29
1.67
0.05
98.45
0.28
Ku-99-SA8
σ (N=42)
38.05
0.35
33.09
0.35
0.03
0.03
0.02
0.01
22.49
0.39
0.03
0.01
7.37
0.12
0.03
0.05
0.01
0.01
101.12
0.45
1.69
0.04
99.43
0.45
S, f
Ku-98-59b
σ (N=2)
38.68
0.26
33.58
0.06
0.17
0.03
0.02
0.01
22.95
1.95
0.03
0.01
7.30
0.07
0.05
0.05
0.01
0.01
102.78
2.09
1.68
0.01
101.09
2.10
S, ne
Ku-99-15a
σ (N=6)
37.87
0.18
33.20
0.59
0.01
0.02
0.04
0.03
23.02
0.52
0.01
0.01
7.40
0.07
0.03
0.06
0.01
0.01
101.62
0.59
1.71
0.03
99.91
0.59
Table 14.5: Average compositions of sodalite (σ standard deviation).
215
unclear what happened to SiO2, released during the sodalitization of albite in all other rock types.
In the investigated samples of the fenitized wallrock lithologies, the carbonatitic breccia and the
late ferrocarbonatite veins, free quartz has never been observed. Two explanations may account
for this observation: (1) SiO2 is incorporated into other silicate phases (i.e. biotite, cancrinite,
muscovite) formed during the fenitization, or (2) SiO2 was removed into higher and not exposed
crustal levels.
14.5 Cancrinite group
Na-Ca feldspathoids, which, based on their optical properties, have been assigned to the
cancrinite group occur sporadically in both the carbonatitic breccia and the fenitized nepheline
syenite, where they formed as an alteration product of albite and nepheline, respectively.
Formulae were calculated on an anhydrous 24-oxygen basis (Table A.5.1.20 in the Appendix
A.5.1). Remarkably, all analysed Na-Ca feldspathoids deviate significantly from the ideal
cancrinite formula (Na,Ca)7-8[Al6Si6O24](CO3,SO4,Cl)1.5-2.0·1-5H2O in containing variable but
generally too low amounts of Na, which apparently increase with decreasing Si (Fig. 14.11).
The
calculated
formulae
of
the
Ca-Na
feldspathoids
range
between
Na4.0Ca0.9K0.2[Al6Si6O24] (nepheline syenite) and Na2.2Ca1.0[Al6Si6.5O24] (carbonatitic breccia).
Regarding their chemical characteristics (e.g. deficiencies of 9-11 wt.% in the total sums of
analyses, not calculating CO2 and H2O; Si/Al ratios of ~ 1:1) the phases resemble selected
members of the zeolite group, such as thomsonite; however, their uniaxial and optically negative
character as well as their granular morphology strongly argue against this possibility. From the
chemically similar mineral nepheline the Ca-Na feldspathoids have been distinguished by optical
properties (stronger birefringence), low totals of analyses (89-91 wt.%) and higher SO3- and Clcontents of up to 0.5 wt.% and up to 0.6 wt.%, respectively.
It can thus be concluded that either (1) the Ca-Na feldspathoids represent a yet unknown,
low-Na member of the cancrinite-group or, more likely, (2) the conspicuous compositions are
caused by submicroscopic intergrowths between cancrinite and at least a second phase, which
contains higher Si and lower Na (such as albite and/or analcite). By now, possibility (2) is the
preferred model, since the apparent negative correlation of Si and Na is quite untypical of
216
feldspathoids, which, on the contrary, are generally characterized by substitution of the type
NaSi ↔ CaAl (Fig. 14.11).
8.0
6.4
6.2
6.0
Al (a.p.f.u.)
Ca+Na+K(a.p.f.u.)
7.0
5.0
4.0
6.0
5.8
5.6
3.0
2.0
5.4
5.4
5.8
6.0
6.3
6.5
5.4
5.6
5.8
Si (a.p.f.u.)
2.5
6.0
6.2
Si (a.p.f.u.)
6.4
6.6
8.0
7.0
2.0
f
Na (a.p. .u.)
Ca (a.p.f.u.)
6.0
1.5
5.0
4.0
1.0
3.0
0.5
2.0
0.0
1
2
3
4
5
Na (a.p.f.u.)
6
7
1.0
5.50
5.75
6.00
Si (a.p.f.u.)
6.25
6.50
Fig. 14.11: Correlation patterns of Si, Na, Ca, K and Al p.f.u. (formula calculation based on 24 oxygens) in the CaNa feldspathoids (black dots). Compositional ranges of members of the cancrinite-vishnevite series (grey shaded
fields), summarized by Deer et al. (1963) are illustrated for comparison
14.6 Analcite
Analcite (general formula Na[AlSi2O6]·H2O has been observed in the fenitized syenite Ku-9859b, where it replaces metasomatically formed albite. Analcite formulae have been calculated on
a basis of 7 oxygens. Due to the small grain sizes of analcite only a few reliable analyses were
obtained, which are listed in Table A.5.1.21 in the Appendix A.5.1.
217
The analysed analcites have an almost ideal composition. The amounts of K2O, CaO and
Na2O are generally near the detection limit. The metasomatic formation of analcite at the
expense of albite can be written as:
Na[AlSi3O8] + H2O → Na[AlSi2O6]·H2O + SiO2
Ab
Anl
(14-3)
in solution
The reaction constrains the interaction of syenite with Si-deficient, Na-rich hydrous
fenitizing fluids, most probably released by the carbonatite melt.
14.7 Aegirine
Clinopyroxene has been analysed in the fenitized ultramafic xenoliths, where it is intimately
associated with magnesio-arfvedsonite and breunnerite. The calculation of the pyroxene formula
followed the method of Papike & Cameron (1980) on the basis of 8 oxygens. For the estimation
of the Fe3+ contents a calculation modus on the basis of 4 cations was chosen. Representative
analyses are listed in Table A.5.1.22 in the Appendix A.5.1.
(Wo,En,Fs)
Ca-Mg-Fe
pyroxenes
UMX
Ku-97-43a
omphacite
jadeite
Jd
aegirineaugite
aegirine
Ac
Fig. 14.12: Compositional variability of clinopyroxene from the fenitized ultramafic xenolith, expressed in the
ternary diagram Jd (NaAlSi2O6) – Ac (NaFe3+Si2O6) – (Wo,En,Fs)((Ca,Mg,Fe)2Si2O6).
Following the classification of Morimoto (1988) all clinopyroxenes of the fenitized
ultramafic xenoliths are aegirine, next to a pure aegirine composition (Fig. 14.12). The Jd
content of the aegirines is generally below 3.3 mol.%, whereas the amounts of (En+Fs+Wo) may
reach up to 5.6 mol.%.
218
The analysed aegirines are most probably not a primary magmatic constituent of the
ultramafic xenoliths, but were formed during fenitization, which additionally led to the
replacement of olivine by magnesite and to the formation of hydrous Fe-Mg silicates (biotite and
amphibole).
14.8 Amphibole
Minor bluish amphibole occurs in both the fenitized anorthosites and in the fenitized ultramafic
xenoliths. The structural formula of amphibole was calculated on an anhydrous 23-oxygen basis
(Table A.5.1.4 in the Appendix A.5.1). For the estimation of minimum and maximum Fe3+
contents a calculation modus on the basis ΣNa=15 and ΣMn=13, respectively, was chosen (Spear
& Kimball, 1984). The average Fe3+ contents were calculated as mean value between the two
extreme compositions, following the method of Papike et al. (1974), modified by Franz &
Häussinger (1990). Amphiboles from both rock suites can be classified as sodic amphiboles in
the sense of Leake et al. (1997).
b) Sodic amphiboles: NaB > 1,50; (Mg + Fe2+ + Mn2+ ) > 2.5;
(AlVI or Fe3+ ) > Mn3+; (Mg or Fe2+ ) > Mn2+
(Na + K)A > 0.50
2+
1,00
Mg/(Mg + Fe )
2+
Mg/(Mg + Fe )
a) Sodic amphiboles: NaB > 1,50; (Mg + Fe2+ + Mn2+) > 2.5;
(AlVI or Fe3+ ) > Mn3+; (Mg or Fe2+ ) > Mn2+
(Na + K)A < 0.50
glaucophane
(AlVI > Fe3+)
fenitized anorthosite
fenitized ultramafic
xenolith
0,75
1,00
eckermannite
VI
3+
(Al > Fe )
nyboite
VI
3+
(Al > Fe )
magnesio-arfvedsonite
VI
3+
(Al < Fe )
ferric-nyboite
VI
3+
(Al < Fe )
ferroeckermannite
(AlVI > Fe3+ )
ferronyboite
(Al VI > Fe 3+)
arfvedsonite
(AlVI < Fe3+ )
ferricferronyboite
(AlVI < Fe3+ )
0,75
magnesio-riebeckite
(AlVI < Fe3+ )
0,50
0,50
ferro-glaucophane
VI
3+
(Al > Fe )
0,25
0,25
riebeckite
VI
3+
(Al < Fe )
0,00
8,0
7,5
7,0
Si p.f.u.
0,00
8,0
7,5
7,0
Si p.f.u.
Fig. 14.13: Compositional variability of sodic amphibole from the fenitized anorthosite (a) and the fenitized
ultramafic xenoliths (b) , expressed in the classification diagram of Leake et al. (1997).
Following Leake et al. (1997) sodic amphiboles from the fenitized anorthosite (Ku-9871), sampled in ~20 cm distance from the contact with a large dyke of the sodalite-rich
carbonatitic breccia, are magnesio-riebeckites (Fig. 14.13a). The Si contents of the amphiboles
are variable (Si p.f.u.: 7.4-7.9), whereas the XMg is relatively constant (XMg: 0.75-0.82).The
219
formation of the magnesio-riebeckites at the expense of unhydrous Fe-Mg silicates (presumably
pyroxene) most probably results from the injection of Na-rich, hydrous and oxidising fluids,
released by the intruding carbonatite magma. The fact, that sodic amphibole is lacking in
anorthositic rocks in direct contact to the carbonatites, which, in contrast, contain abundant
biotite, is taken as evidence for a chemical zonation of the fenitic aureoles with respect to Na and
K.
Sodic amphiboles, which together with aegirine and magnesite, constitute the finegrained matrix of the fenitized ultramafic xenoliths, are magnesio-arfvedsonites (Fig 14.13b).
The amphiboles are characterized by high Si (Si p.f.u.: 7.7-8.0) and Na contents (Na p.f.u.: 2.12.6) and a high XMg of 0.81-0.89. The intimate association of magnesio-arfvedsonite with
secondary magnesite suggests a secondary formation of the amphibole, caused by interactions
between the xenoliths and Na- and H2O-rich fluids.
14.9 White mica
Microprobe analyses were carried out for muscovite from the fenitized syenite, the fenitized
nepheline syenite and the carbonatitic breccia. Muscovite formulae were calculated on an
anhydrous basis of 22 oxygens. Representative analyses are depicted in Table A.5.1.23 in the
Appendix A.5.1.
As shown in Fig. 14.14, the compositional variation of the analyzed white micas from the
fenitized nepheline syenite and the carbonatitic breccia is in excellent agreement with a
Tschermak's substitution mechanism (AlIV+AlIV = R2++Si4+), even though micas from the
carbonatitic breccia generally plot below this line, indicating an increased substitution of Fe3+ for
AlVI. In clear contrast, the trend displayed by data points of white mica of the fenitized syenite
rather implies edenite exchange (NaAlIVo-1Si-1; "o" = vacancy).
The analysed white micas display significant differences in their composition depending
on the respective rock type. White micas of the carbonatitic breccia (Si p.f.u. 5.91-6.07, Fe p.f.u.:
0.02-0.31, Na p.f.u.: 0.02-0.34) are generally characterized by higher Na and Fe contents when
compared to those of the fenitized nepheline syenite (Si p.f.u. 5.93-5.99, Fe p.f.u.: 0.01-0.09, Na
p.f.u.: 0.03-0.0.09) and the fenitized syenite (Si p.f.u. 5.74-6.06, Fe p.f.u.: 0.02-0.19, Na p.f.u.:
220
0.02-0.0.09). The white micas are characterized by a high XFe, ranging between 0.67 and 0.95
(carbonatitic breccia), 0.55 and 1.00 (fenitized syenite) and 0.83-1.00 (fenitized nepheline
syenite). Chlorine and fluorine contents of all white micas analysed are generally below the
detection limit.
6.1
Si p.f.u.
6.0
SiR 2+
A
5.9
5.8
5.7
5.6
5.7
l IV-
1
Al IV
-1
oS
iN
a-
fenitized syenite
1
Al
IV
-1
CB,so
Ku-98-58a
Ku-98-131
CB,l
Ku-98-56
5.8
5.9
6.0
6.1
6.2
6.3
Al p.f.u.
Fig. 14.14: Variation of Si versus Al in white micas of the fenitized syenite, the fenitized nepheline syenite and the
Following the classification of Schliestedt (1980) the composition of the analysed white
micas is muscovite-dominated, with variable amounts of the paragonite and the celadonite
components, which, however, are more prominent in white micas of the carbonatitic breccia
(Fig. 14.15). The differences in the Na and Fe-contents between muscovites from the
carbonatitic breccia and those of the fenitized syenite and the fenitized nepheline syenite may be
explained by their different crystallization histories. Judging from its textural position, muscovite
from the nepheline syenite and the syenite crystallized synchronously with the Na-fractionating
phases sodalite and albite, respectively. In clear contrast, muscovite from the carbonatitic breccia
clearly post-dates albite and sodalite and thus may have incorporated the maximum possible
amount of Na available in the respective stage of metasomatism.
221
In one sample of the layered carbonatitic breccia (Ku-98-14), white mica has a somewhat
unusual composition, deviating significantly from that of muscovite. The micas contain large
amounts of Na2O (1.16-1.19 a.p.f.u.) and minor Ba (0.1 a.p.f.u.), both substituting K (0.67-0.75
a.p.f.u.). Remarkably, these micas plot in the compositional gap of muscovite-paragonite solid
solution in the ternary diagram paragonite-muscovite-celadonite after Schliestedt (1980; Fig.
14.15). The conspicuous apparent composition of the white mica could be caused by
submicroscopic changes in its interlayer occupations, that are predominantly occupied by either
Na or K. This interpretation, however, has to be examined with the help of detailed TEM studies.
Pg
paragonite
fenitized syenite
CB,so
Ku-98-58a
Ku-98-131
composition gap
ado
sco
v
ite
cel
CB,l
Ku-98-14
Ku-98-56
mu
e
nit
phengite
Ms
Ce
Pg
Pg
80
90
Ms
90
80
te
mu
sco
vi
mu
sc o
v it
e
80
90
phengite
70
60
Ce
Ms
90
phengite
80
70
60
Ce
Fig. 14.15: Classification of metasomatically formed white micas in the study area, expressed in the paragonitemuscovite-celadonite diagram according to Schliestedt (1980).
14.10 Biotite
Biotite is a major constituent of the fenitized anorthosite, the carbonatitic breccia and the
fenitized ultramafic xenoliths. Moreover biotite occurs as a minor orthomagmatic phase in
222
nepheline syenite. Biotite formulae were calculated on an anhydrous 22-oxygen basis.
Representative analyses are depicted in Table A.5.1.5 in the Appendix A.5.1.
The analysed biotites display considerable compositional variations, between both rock
types and individual samples and were thus subdivided into the following seven groups:
•
metasomatically formed biotite from sodalite-biotite-assemblages of xenoliths of the
fenitized anorthosite incorporated by the sodalite-rich carbonatitic breccia (Bt1),
•
magmatic biotite in intimate association with apatite in samples of the sodalite-rich
carbonatitic breccia (Bt2),
•
magmatic biotite associated with apatite in sodalite-free samples of the layered
carbonatitic breccia (Bt3)
•
metasomatically formed biotite from the fenitized syenite (Bt4)
•
orthomagmatic biotite from the nepheline syenite (Bt5) and
•
metasomatically formed biotite from biotite-ankerite patches and schlieren of the
fenitized ultramafic xenoliths (Bt6)
•
orthomagmatic biotite from the K-feldspar-biotite-ilmenite-calcite rock (Bt7).
In assemblage (1), where biotite from the carbonatitic breccia coexists with sodalite, XMg
clusters around 0.43 (XMg: 0.37-0.60) whereas the biotite in assemblage (2) and (3) exhibits
average XMg values of 0.49 (0.44-0.57) and 0.44 (0.37-0.52), respectively. All biotites are
characterized by high Ti contents (Bt1: Ti p.f.u.: 0.23-0.32; Bt2: Ti p.f.u.: 0.12-0.38 and Bt3: Ti
p.f.u.: 0.16-0.41) and minor oscillatory to concentric zoning, involving a core preference for Al,
whereas the rim trends to higher Si, pointing to a progressively lower availability of Al in the
carbonatite magma. Decreases in Al are accompanied by increases of the XMg, suggesting that Fe
was consumed during the synchronous crystallization of magnetite. Following the modified
classification diagram of Foster (1960, in Tröger, 1982) the biotites are meroxenes to
lepidomelanes (Fig. 14.16). The fact, that most of the magmatic biotites, coexisting with apatite
in the carbonate-rich zones, plot slightly below the meroxene and lepidomelane fields is taken as
evidence for elevated Fe3+ contents of the micas when compared to those coexisting with sodalite
in the xenolith-rich zones. This interpretation is confirmed by Al-deficiencies of the tetrahedral
position of the biotites which may be compensated by either Ti or Fe3+.
Metasomatically formed biotite from the fenitized syenite (Bt4) is lepidomelane with
patchy zonation (Ti p.f.u.: 0.15-0.34; Fig. 14.17a). The XMg attains comparably low values of
223
0.26-0.39, reflecting the Fe-rich composition of the primary magmatic Fe-Mg silicates (pyroxene
and amphibole), which were replaced by biotite during the metasomatic alteration of the
syenites. Magmatic biotites from the nepheline syenite (Bt5) are unzoned meroxenes to
lepidomelanes (XMg: 0.50-0.57; Ti p.f.u.: 0.23-0.26; Fig. 14.17b) whereas magmatic biotites
from the K-feldspar-biotite-ilmenite-calcite rock (Bt7) are unzoned meroxenes (XMg of 0.610.63; Ti p.f.u.: 0.31-0.36; Fig. 14.17c).
R3+ = AlVI+Ti4+
R3+ = AlVI+Ti4+
CB,so
Ku-97-15
(N=19)
CB,so
Ku-98-80
(N=21)
sidero- siderophyllite
phyllite
phyllite
lepidomelane
eastonite
lepidomelane
eastonite
meroxene
phlogopite
phlogopite
a
tot
Mg
phlogopite
meroxene
Fe +Mn
annite
b
tot
Mg
phlogopite
R3+ = AlVI+Ti4+
Fe +Mn
annite
3+
VI
R = Al +Ti
CB,so
Ku-98-131
(N=231)
4+
CB,so
Ku-99-SA8
(N=147)
phyllite
phyllite
lepidomelane
eastonite
lepidomelane
eastonite
meroxene
phlogopite
phlogopite
c
Fetot+Mn
annite
Mg
phlogopite
3+
VI
meroxene
d
Fetot+Mn
annite
Mg
phlogopite
4+
3+
VI
R = Al +Ti
R = Al +Ti
CB,l
Ku-98-14
(N=48)
CB,l
Ku-98-48
(N=52)
4+
phyllite
phyllite
lepidomelane
eastonite
lepidomelane
eastonite
meroxene
phlogopite
Mg
phlogopite
meroxene
phlogopite
e
Fetot+Mn
annite
Mg
phlogopite
f
tot
Fe +Mn
annite
Fig. 14.16: Biotite classification in the modified (Fe2+=Fetot) triangle R3+(AlVI+Ti)-Mg-(Fetot+Mn) after Foster
(1960, in Tröger, 1982). a) and b) Metasomatically formed biotite (Bt1) from the sodalite-rich carbonatitic breccia.
c) and d) Magmatic biotite (Bt2) from the sodalite-rich carbonatitic breccia. e) and f) Magmatic biotite (Bt3) from
the layered carbonatitic breccia.
224
The highest overall Mg and lowest Ti contents are displayed by biotite from the fenitized
ultramafic xenoliths (Bt6; XMg: 0.67-0.77; Ti p.f.u.: 0.03-0.06), in good accordance with the
Mg-rich compositions of coexisting ankerites/dolomites. Due to the low Al and Ti amounts of
the biotites, there was still some deficiency left in the tetrahedral site (Si+AlIV+TiIV) of the
calculated biotite formulae. Thus an appropriate amount of Fe was recast as Fe3+ in order to
complete the tetrahedral site occupancy, following the equation Fe3+=8-Si-AlIV-TiIV. All
analysed biotites from different textural positions such as biotite in ankerite-biotite patches in the
xenoliths and biotite in reaction rims between the xenoliths and the enclosing carbonatitic
breccia are meroxenes in the sense of Foster (1960, in Tröger, 1982) and show no significant
differences in the chemical compositions (Fig. 14.17d). It can thus be concluded that the
formation of all biotites of the ultramafic xenoliths was caused by the interaction between the
ultramafic xenolith and the transporting carbonatite magma under high oxygen fugacities.
3+
VI
4+
3+
VI
4+
R = Al +Ti
R = Al +Ti
S,f
Ku-98-59b
(N=223)
S,ne
Ku-99-15a
(N=25)
phyllite
phyllite
lepidomelane
eastonite
lepidomelane
eastonite
meroxene
phlogopite
Mg
phlogopite
phlogopite
a
tot
3+
VI
Fe +Mn
annite
4+
R = Al +Ti
meroxene
Mg
phlogopite
L
Ku-98-25
(N=10)
b
tot
3+
VI
4+
3+
R = Al +Ti +Fe
UMX
Ku-97-43a
(N=81)
phyllite
phyllite
lepidomelane
eastonite
lepidomelane
eastonite
meroxene
phlogopite
Mg
phlogopite
Fe +Mn
annite
meroxene
phlogopite
c
tot
Fe +Mn
annite
Mg
phlogopite
d
2+
Fe +Mn
annite
Fig. 14.17: Biotite classification in the modified triangle R3+(AlVI+Ti)-Mg-(Fetot+Mn) (a-c) and the triangle
R3+(AlVI+Ti+Fe3+)-Mg-(Fe2++Mn) (d) after Foster (1960, in Tröger, 1982). a) Metasomatically formed biotite (Bt4)
from the fenitized syenite. b) Magmatic biotite (Bt5) from the fenitized nepheline syenite. c) Magmatic biotite (Bt7)
from the K-feldspar-biotite-ilmenite-calcite rock. (Bt7) d) Metasomatically formed biotite (Bt6) from the ultramafic
xenoliths (Fe3+ contents recalculated).
In the discrimination diagram MgO versus TiO2 (Fig. 14.18), both the orthomagmatic
biotites of the carbonatitic breccia (Bt2 and Bt3) and the metasomatically formed biotites of the
225
ultramafic xenoliths (Bt6) mostly plot within the field of biotite from carbonatites as defined by
Gaspar & Wyllie (1987). Metasomatic biotites from the carbonatitic breccia (Bt1)
compositionally resemble those of magmatic origin. In clear contrast, metasomatically formed
biotites from the fenitized syenites deviate from this composition in displaying lower Ti and Mg
contents, reflecting the Fe-rich and Ti-poor whole-rock composition of the syenites (Fig. 14.18).
Distinct trends are also displayed by the Si/Al-ratios of the different biotite subgroups
(Fig. 14.19). Magmatic biotite of the K-feldspar-biotite-ilmenite-calcite rock (Bt7) plots near the
ideal biotite composition. In contrast, both, Bt1 and Bt4 have Al in excess of that necessary to
fill the tetrahedral site in addition to Si. The biotites exhibit a scattered distribution in the Si
versus Al diagram but mainly plot within the compositional field of late-magmatic biotite of the
4.5
fenitized syenite
fenitized ultramafic
xenolith
CB,so
Ku-97-15
Ku-98-80
Ku-98-131
Ku-99-SA8
CB,l
Ku-98-14
Ku-98-48
4.0
TiO2 (wt.%)
3.5
3.0
2.5
2.0
1.5
biotite from carbonatites
(Gaspar & Wyllie, 1987)
1.0
0.5
0.0
0
5
10
15
MgO (wt.%)
20
25
30
Fig. 14.18: Compositional range of magmatic and metasomatically formed biotite from the carbonatitic breccia, the
fenitized syenite and the fenitized ultramafic xenoliths, expressed in the diagram MgO versus TiO2 after Gaspar &
Wyllie (1987).
unaltered anorthosite lithologies and in the range of magmatic biotite of the nepheline syenite
(Bt4). This is taken as supporting evidence, that Bt1 and Bt4 were formed by interaction between
the carbonatite magma and the bordering anorthosites and syenites, respectively, with the latter
two providing the necessary Al. Bt2 and Bt3, which both crystallized from the carbonatite melt
display a distinct trend, reaching from high Al values towards Al deficit compared to ideal
biotite, evidencing that either (1) the carbonatite melt contained only minor Al amounts, that
226
were consumed during the biotite formation or that (2) the biotites crystallized under
progressively increasing fO2, causing substitution of Fe3+ for Al. Low Al contents of Bt1, Bt2,
Bt3 and Bt4 are accompanied by high Mg and/or low Na and high Si, suggesting the most
important substitution mechanism could be AlVI+AlIV=MgVI+2SiIV (tschermak), and, far less
frequently, Na+AlIV=o+2SiIV (edenite scheme). In Bt2 and Bt3 Fe3+ increases with decreasing
AlIV, suggesting an additional substitution scheme of Fe3+↔AlIV. Bt6, coexisting with
ankerite/dolomite in the fenitized ultramafic xenoliths, is Al- and Ti-poor, hence strongly
suggesting substitution of Fe3+ for Al. The low Al contents of these micas most probably result
from the generally Al-poor composition of both the ultramafic xenoliths and the transporting
carbonatite magma.
3.10
fenitized syenite
fenitized ultrabasic
xenolith
Kfs-Bt-Ilm-Cc rock
CB,so
Ku-97-15
Ku-98-80
Ku-98-131
Ku-99-SA8
CB,l
Ku-98-14
Ku-98-48
3.05
ideal
phlogopite
3.00
tetraferriphlogopite
2.90
Si +
2.85
Al
=
4
Si a.p.f.u.
2.95
2.80
2.75
compositional
range of biotite
in anorthosite
2.70
2.65
2.60
0.90
1.00
1.10
1.20
1.30
1.40
1.50
1.60
1.70
1.80
Al a.p.f.u.
Fig. 14.19: Variation of Si versus Al in biotites of the study area.
The highest but most variable Cl contents of up to 0.24 wt.% are displayed by Bt4 of the
fenitized syenite (XMg: 0.26-0.39; Fig. 14.20), which at the same time contains the lowest F
contents (0.4-1.5 wt.%). In clear contrast, orthomagmatic Bt2 and Bt3 in assemblage with apatite
contain Cl amounts at or below the detection limit (up to 0.11 wt.%; Fig. 14.20) corresponding to
F contents (0.3-2.7 wt.%) and a higher XMg (0.37-0.57). Lowest Cl contents of 0–0.05 wt.% are
227
revealed by the Mg-rich biotites (XMg: 0.62-0.73) of the fenitized ultramafic xenoliths (Fig.
14.20), which, however, are characterized by the highest F contents (1.5-3.8 wt.%).
The results are in good agreement with experimental studies of Munoz & Swenson
(1981) and Volfinger et al. (1985), who suggest a preferential incorporation of Cl into Fe-rich
biotites, even though the predicted trend of “Mg-Cl-avoidance” is not very pronounced in the
rock-samples investigated. Especially Bt4 in sample Ku-98-59b, has Cl contents, which are
much more variable than the corresponding XMg. Therefore, it appears that Cl incorporation in
biotite is not solely governed by the Mg-Cl-avoidance but may additionally result from the
interaction of biotite with compositionally changing metasomatic fluids. Following this, fluidflow outlasted the biotite formation.
0.16
fenitized syenite
fenitized ultramafic xenoliths
0.14
CB,so
Ku-98-131
Ku-99-SA8
CB,l
Ku-98-14
Ku-98-48
Cl (wt.% )
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0.20
0.30
0.40
0.50
0.60
0.70
0.80
XMg
Fig. 14.20: Chlorine contents (wt.%) of magmatic and metasomatic biotites plotted against the respective mole
fraction of Mg.
14.11 Apatite
Apatite is a minor constituent of both the layered and the sodalite-rich carbonatitic breccia.
Apatite compositions were investigated in great detail, since the mineral may incorporate F, Cl
228
and OH and thus gives additional information on the composition of the carbonatite magma.
Representative analyses are presented in Table A.5.1.24 in the Appendix A.5.1. Mineral
formulae were calculated on a 25-oxygen basis. Apatites from both the layered carbonatitic
breccia and the sodalite-rich carbonatitic breccia contain significant amounts of F (1.3-5.1
wt.%) and Sr (1.1-2.5 wt.%) and can thus be classified as strontian fluorapatites.
The F-contents of apatite increase from an average value of 2.80 wt.% (1.3-4.8 wt.%) in
sodalite-rich samples to 3.64 wt.% (2.2-5.1 wt.%) in the layered carbonatitic breccia. In the same
sequence of samples the concentrations of Na2O (0.33-1.09 wt.%) and BaO (0-0.14 wt.%) and Sr
increase. Most of the investigated apatites exhibit only minor amounts of Cl, which, however,
may reach values of up to 0.31 wt.% in single apatites, corresponding to 0.1 Cl atoms per
formula unit.
1
CB,so
Ku-99-SA8
Ku-98-131
CB,l
granite
pegmatite
Ku-98-14
Mn (wt.%)
Ku-98-48
0.1
carbonatite
phosphorite
0.01
skarn
Sr-enrichment
0.001
0.001
0.01
0.1
Sr (wt.%)
1
10
Fig. 14.21: Mn-Sr fields of apatite from skarn, phosphorite, granite pegmatite and carbonatite from world-wide
localities (compiled by Hogarth, 1989). Note that all apatites from the Swartbooisdrif carbonatites plot in the
carbonatite field.
Trace element contents of FeO (<0.54 wt.%), MnO (<0.16 wt.%) and MgO (0.0 wt.%)
are generally extremely low. Consistently low total sums of analyses of apatite from one sample
of the layered carbonatitic breccia (Ku-98-14) might be explained by the additional incorporation
of carbon. However, the exact substitution scheme remains unclear, since chemical analyses of
229
OH- and [CO3]2- are still wanting. In the Sr versus Mn diagram (Fig. 14.21) all analysed apatites
from the carbonatitic breccia fall in the compositional range of apatites from carbonatites.
14.12 Pyrochlore
Medium- to very fine-grained and optically zoned pyrochlore occurs as an accessory phase in
both the late ferrocarbonatite veins and the carbonatitic breccia. However, the small pyrochlore
grains from the carbonatitic breccia are completely altered to fine-grained intergrowths of
secondary phases and hence attained unreasonable and highly variable EMP analyses. Formula
properties of pyrochlore from the ferrocarbonatite veins have been calculated on the basis of
total occupation on the B-site of pyrochlore (Nb + Si + Ti = 2), according to the general formula
A-ion characterization
Na + Ca but no other A-atom
Pyrochlore subgroup
Microlite subgroup
Betafite subgroup
Nb+Ta > 2Ti; Nb > Ta
Nb+Ta > 2Ti; Ta ≥ Nb
2Ti ≥ Nb+Ta
PYROCHLORE
microlite
calciobetafite
> 20% total A-atoms
One ore more A-atoms,
other than Na or Ca,
> 20 % total A-atoms
Species named by most
abundant A-atom, other
than Na or Ca
K
Cs
Sn
Ba
Sr
kalipyrochlore
REE (ΣCe > ΣY)
REE (ΣY > ΣCe)
Bi
Sb
Pb
U
Th
ceriopyrochlore
yttropyrochlore
bariopyrochlore
(un-named)
cestibtantite
stannomicrolite
bariomicrolite
yttrobetafite
bismutomicrolite
plumbopyrochlore
uranpyrochlore
(un-named)
plumbomicrolite
uranmicrolite
stibiobetafite
plumbobetafite
betafite
Table 14.6: Classification of the pyrochlore group after Bonshtedt-Kupletskaya (1966; in Hogarth, 1989) and
Hogarth (1977) (REE = lanthanides + Y, ΣCe = (La→Eu), ΣY = (Gd→Lu) + Y, pyrochlore species found in the
Swartbooisdrif carbonatite is in capitals).
(A)2-mB2O6(O,OH,F)1-n pH2O (Hogarth, 1977). However, with values of 1.64-1.88, the totals of
the remaining cations in the A-site (Fe,Ce,Ca,Sr,Ba,Na) of the analysed pyrochlores are
distinctly lower than those given by the ideal formula (Ca,Na)2Nb2O6(OH,F)2. These deficiencies
in the A-site is a common feature of pyrochlores from carbonatites and are generally attributed to
either (1) vacancies in the A-site or (2) strong hydration. Representative analyses of pyrochlore
are given in Table A.5.1.25 in the Appendix A.5.1.
230
The characterization of pyrochlores followed the classification scheme of Bonshtedt-
Kupletskaya (1966; in Hogarth, 1989) and Hogarth (1977), which is based on the B-site and Asite characterization of pyrochlore (Table 14.6). Following this classification scheme,
pyrochlores are first subdivided into three subgroups, i.e. (1) pyrochlore Nb+Ta > 2Ti, Nb > Ta,
(2) microlite Nb+Ta > 2Ti, Ta ≥ Nb and (3) betafite 2Ti ≥ Nb+Ta. Subsequently, subspecies of
the subgroups are defined as (i) Na and Ca but no other A-atom exceeding 20 at.% of the totals
of A-atoms and (ii) one or more A-atoms other than Ca and Na exceeding 20 at.%. With an
average formula of Ca0.83Na0.85Sr0.05Ce0.03Fe0.01Nb1.90Ti0.08 Si0.02O6 (OH,F)2, all investigated
pyrochlores plot in the field of the pyrochlore subspecies.
EMP analyses generally reveal high SrO and Na2O contents of 0.96-2.24 wt.% and 6.547.98 wt.%, respectively, whereas the Ta contents are < 0.7 wt.%. The fine-scaled oscillatory
zonation of pyrochlore, observed microscopically, is also displayed by its mineral chemistry. All
analysed grains show a concentric zoning pattern with a core preference for Ca and Ti, whereas
Si, Fe, Ba, Sr and Ce trend to rim.
14.13 Fe-Ti Oxides
Ilmenite, hematite, rutile and magnetite are ubiquitous accessory constituents of most samples
investigated. The calculation of the ilmenite and hematite formula was based on 6 oxygens
whereas magnetite was calculated on a 4-oxygen basis. For the estimation of the Fe3+ contents of
ilmenite/hematite and magnetite a calculation modus on the basis of 4 and 3 cations,
respectively, was chosen. Rutile was calculated on a 2-oxygen basis. Representative analyses of
ilmenite, magnetite and hematite are given in Tables A.5.1.7, A.5.1.A and A.5.1.9 in the
Appendix A.5.1. Rutile analyses are listed in Table A.5.1.26 in the Appendix A.5.1.
14.13.1 Ilmenite
Ilmenite occurs in several different textural positions and rock samples, i.e. (1) as oxy-exsolution
lamellae in magnetite from the fenitized anorthosites, (2) as anhedral, heavily tectonised grains
in the fenitized syenite, (3) as late-magmatic, anhedral grains in the K-feldspar-biotite-ilmenitecalcite rock and (4) as small, irregularly bounded grains in the carbonatitic breccia.
231
The MnO contents of ilmenite oxy-exsolution lamellae in magnetite from the fenitized
anorthosite are generally low, ranging from 0.6-0.7 wt.% The hematite contents of these
ilmenite types fall in a range of 0.4-0.6 mol.%.
In contrast, discrete grains of ilmenite in the fenitized syenite are characterized by
comparably high MnO amounts of 5.37-6.04 wt.% and an extremely low hematite content of
<0.01 mol.%, in good accordance with the ilmenite compositions from unaltered syenites.
In the K-feldspar-biotite-ilmenite-calcite rock, ilmenite occurs as discrete grains in the
interstices between plagioclase and biotite crystals. EMP analyses reveal ilmenite compositions
with minor amounts of the Mn end-member pyrophanite (3.2-4.1 mol.%; MnO: 1.5-1.9 wt.%)
and low but quite variable contents of hematite (0.1-1.3 mol.%).
Anhedral ilmenite occurs in rare samples of the carbonatitic breccia, where the oxide
generally forms single grains, bearing abundant inclusions of matrix minerals. In places, ilmenite
is marginally replaced by rutile or altered to titanohematite along cracks. EMP analyses along
various profiles reveal high hematite contents of 1.6-4.3 mol.% and low MnO values of 0.8-1.1
wt.%.
14.13.2 Magnetite
Magnetite is a common accessory phase in the fenitized anorthosites, the fenitized syenites, the
carbonatitic breccia, the late ferrocarbonatite veins and the fenitized ultramafic xenoliths.
Magnetite from the fenitized anorthosites occurs as intercumulus grains with ilmeniteexsolution lamellae. However, along the lamellar systems and along cracks magnetite is mostly
completely altered to hematite replaced by ankerite, whereby ilmenite exsolution lamellae are
preserved. Thus only few reliable analyses of magnetite were obtained, revealing TiO2 contents
of 0.02-0.13, which are in the range of those obtained for similar magnetites from the unaltered
anorthositic rocks.
Magnetite from the fenitized syenite occurs as large anhedral grains, displaying embayed
grain boundaries and containing abundant albite inclusions. The homogeneous magnetites attain
Ti contents of 0.0-0.3 wt.%, in good accordance with the Ti values obtained for magnetites of
unaltered syenites.
232
In the carbonatitic breccia, magnetite occurs in two different textural positions, i.e. (1) as
inclusion-free magnetite dispersed in the matrix and in places forming large grains of up to 4 mm
in size and (2) as irregularly bounded grains of up to 2 cm in diameter, which frequently display
inclusions of matrix minerals. However, magnetites from both textural position are almost
identical in composition, containing extremely low Ti contents of < 0.9 wt.%, suggesting
subsolidus re-equilibration of all oxides during the final stages of metasomatism. In the late
ferrocarbonatite veins, angular or irregularly shaped magnetite grains are evenly distributed in
the ankerite matrix. Analysis profiles across magnetite reveal extremely low values of TiO2
(0.00-0.08 wt.%) and MnO (0.00-0.04 wt.%), both slightly decreasing from core towards the rim.
Magnetite is a minor constituent of the fenitized ultramafic xenoliths, where the oxide is
intimately associated with ankerite/dolomite and biotite in irregular patches. Magnetite was
found to be virtually pure FeFe2O4, containing minor amounts of Ti (0.06-0.11wt.%) in the range
of magnetite from the late ferrocarbonatite veins.
14.13.3 Hematite
Hematite is a common alteration product of both magnetite and ilmenite in samples of the
carbonatitic breccia. Depending on their textural position, the hematites display significant
differences in their chemical composition. When formed by the martitisation of magnetite,
hematite was found to be virtually pure Fe2O3 (TiO2 < 0.1 wt.%), whereas hematite replacing
ilmenite attains distinctly higher TiO2 amounts of 2.1-7.0 wt.%. Individual grains of titanohematite may be zoned with the Ti-contents increasing from core towards the rim. Cr2O3, MgO
and MnO are generally below the detection limit.
14.13.4 Rutile
Rutile is present in various textures in the fenitized anorthosites, the carbonatitic breccia and the
late ferrocarbonatite veins, i.e. (1) as large, anhedral grains of rutile, either isolated or in contact
with magnetite from both the ferrocarbonatite veins and the carbonatitic breccia, (2) as finegrained overgrowths of rutile and carbonate, on magnetite and (3) in intergrowths with hematite,
replacing magnetite twin lamellae and/or ilmenite oxy-exsolution lamellae in magnetite.
233
Rutile from the carbonatitic breccia and the fenitized anorthosites commonly attains
high Fe2O3 contents, ranging between 0.6-2.2 wt.%, but displaying no significant differences
with respect to their textural positions.
Rutile in contact to magnetite and pyrochlore in the late ferrocarbonatite veins reveals
overall high Fe2O3 amounts of 2.2-3.1 wt.%. Low total sums of oxides of the analysed rutiles (<
97.5 wt.%), suggest that the Ti-oxides may additionally contain significant amounts of Nb.
14.14 Sulphides
Sulphides occur as accessory phases in the fenitized anorthosites, the fenitized syenites, the
carbonatitic breccia and the K-feldspar-biotite-ilmenite-calcite rock. It has to be mentioned
however, that the sulphides are quite heterogeneously distributed within the respective rock
types, and are thus lacking in most of the samples investigated.
14.14.1 Pyrite
Pyrite is a common accessory phase of the fenitized anorthosites, the fenitized syenites and the
carbonatitic breccia. In the latter, the sulphides are concentrated in samples which presumably
suffered autometasomatism by late-stage, highly evolved fenitizing fluids, such as the REE-rich
carbonatitic breccia. In addition, dispersed pyrite grains were observed in the K-feldspar-biotiteilmenite-calcite rock. Representative pyrite analyses are listed in Table A.5.1.27 in the Appendix
A.5.1. Compositional ranges of the analysed pyrites are illustrated in the ternary diagram CoS2NiS2-FeS2 after Nickel (1970; Fig. 14.22).
Pyrite from the fenitized anorthosite occurs in two different textural positions, i.e. as (1)
strongly corroded, euhedral grains in contact to magnetite, overgrown by (2) pyrite of a second
generation. The latter may also occur as small, subhedral grains, dispersed in the rocks. Pyrite of
the first generation reveal Co contents near the detection limit, whereas the Ni contents range
between 0.0-1.6 wt.%, with the Ni increasing from core towards the rim (Fig. 14.22a). In clear
contrast, texturally late pyrite is characterized by Ni contents below the detection limit and Cocontents varying between 0.0 wt.% (core) and 3.5 wt.% (rim; Fig. 14.22b). The formation of the
234
pyrites is most probably linked to the sulphidation of orthomagmatic magnetite, following the
injection of the fenitizing carbonatite fluids.
CoS2
50
50
NiS2
FeS2
50
CoS2
CoS2
<
<
core
rim
core
rim
90
90
95
95
a
NiS2
<
b
90
FeS2
95
NiS2
<
90
CoS2
CoS2
<
<
Ku-98-56
Ku-98-130b
Kfs-Bt-Ilm-Cal rock
Ku-98-25
90
90
95
95
c
NiS2
<
90
FeS2
95
d
95
FeS2
NiS2
<
90
95
FeS2
Fig. 14.22: Compositional variation of pyrite from the studied samples, expressed in the ternary diagram CoS2NiS2-FeS2 after Nickel (1970). a) First pyrite generation of the fenitized anorthosites. b) Second pyrite generation of
the fenitized anorthosites. c) Pyrite from the carbonatitic breccia. d) Late-magmatic pyrite from the K-feldsparbiotite-ilmenite-calcite rock.
Two pyrite generations are distinguished in samples of the carbonatitic breccia. A first
generation of pyrite forms small, subhedral grains, which are dispersed in the carbonatite matrix
(Ku-98-56). These pyrites attain high Ni contents of 2.5-2.7 wt.% and Co contents below the
detection limit (Fig. 14.22c). In contrast, pyrite of the second generation is localised in sulphide
235
patches and veins, where it is intimately associated by millerite and chalcopyrite (Ku-98-130b).
The Ni values of these late-stage pyrites are generally < 0.1 wt.% whereas Co ranges between
0.1 wt.% and 2.3 wt.% (Fig. 14.22c). The evolutionary trend of the pyrites towards higher Co
and lower Ni contents is in excellent accordance with the findings for the two pyrite generations
from the fenitized anorthosites, hence suggesting that the composition of the fenitizing fluids
changed during metasomatism. Following this, early fenitizing fluids must have transported
certain amounts of Ni, whereas Co dominates over Ni in the late-stage fluids.
Subhedral pyrite grains of the K-feldspar-biotite-ilmenite-calcite rock are found to be
virtually pure FeS2. Both the Co and the Ni contents are generally low, ranging between 0.0 and
0.5 wt.% and 0.0 and 0.2 wt.%, respectively (Fig. 14.22d). Judging from their textural position
and their chemical composition, these pyrites are suggested to represent a late-magmatic
crystallization product of the K-feldspar-biotite-ilmenite-calcite rocks, unrelated to metasomatic
processes.
14.14.2 Chalcopyrite
Chalcopyrite is closely associated with pyrite in the fenitized anorthosite. In the carbonatitic
breccia, chalcopyrite either occurs as small dispersed grains or localised in sulphide veins and
patches. Representative analyses of chalcopyrite are depicted in Table A.5.1.28 in the Appendix
A.5.1.
Ni and Co contents of chalcopyrites from the fenitized anorthosite are generally below
the detection limit, irrespective of their textural position. The same holds true for late-stage
chalcopyrite of the carbonatitic breccia, which is intimately intergrown with millerite. In clear
contrast, dispersed grains of early chalcopyrite from the carbonatitic breccia reveal slightly
higher Ni contents of up to 0.12 wt.%, hence supporting the interpretation that early fenitizing
fluids are richer in Ni than their late-stage counterparts.
14.14.3 Pyrrhotite
Pyrrhotite was observed in only one sample of the carbonatitic breccia (Ku-97-19; von
Seckendorff & Drüppel, 1999; von Seckendorff et al., 2000). In this rock, pyrrhotite and pyrite
236
form either single grains or irregular, layer-parallel patches. Representative pyrrhotite analyses
are compiled in Table A.5.1.29 in the Appendix A.5.1.
EMP analyses revealed Ni contents of 0.75 wt.% for a pyrrhotite inclusion in pyrite,
whereas matrix pyrrhotite, bearing abundant pentlandite lamellae, is virtually Ni–free (Ni < 0.15
wt.%). The composition of the investigated pyrrhotites ranges between Fe0.97S and FeS.
14.14.4 Pentlandite
Pentlandite occurs as lamellae in pyrrhotite of one sample of the carbonatitic breccia (Ku-97-19;
von Seckendorff & Drüppel, 1999; von Seckendorff et al., 2000) and as small, isolated grains in
the fenitized anorthosite. The formula of pentlandite has been calculated on an 8-sulphur basis.
Representative analyses are listed in Table A.5.1.30 in the Appendix A.5.1.
Pentlandite of the fenitized anorthosite has a rather homogeneous composition of
Fe3.5Ni4.8Co0.3S8, whereas the compositional range of pentlandite lamellae in pyrrhotite from the
carbonatitic breccia can be described as Fe4.5-5.3Ni3.3-4.4S8, with Co contents below the detection
limit. Higher Fe and lower Ni contents of the latter pentlandites, when compared to those of the
fenitized anorthosites, are most probably related to the Fe-rich and Ni-poor composition of the
pyrrhotite host.
14.14.5 Bornite
Small bornite grains were observed in one sample of the carbonatitic breccia (Ku-97-15d),
where the sulphide is associated by chalcocite-digenite (von Seckendorff & Drüppel, 1999; von
Seckendorff et al., 2000). The analysed bornite was found to be virtually pure Cu5FeS4 (Table
A.5.1.31 in the Appendix A.5.1).
14.14.6 Chalcocite-digenite
Members of the chalcocite-digenite group have been observed in only one sample of the
carbonatitic breccia (Ku-97-15d), either as single grains or, in association with bornite, as
irregular grain aggregates (von Seckendorff & Drüppel, 1999; von Seckendorff et al., 2000).
237
EMP analyses of chalcocite-digenite (Table A.5.1.32 in the Appendix A.5.1) yielded two
chemically distinct compositional groups with different Fe-contents. The Fe-poor group is nearbinary, generally with compositions of Cu1.91S - Cu1.94S and Fe contents of 0.05-0.07 wt. %,
whereas the second group reveals Fe values of 1.2-3 wt. %. The data points do not plot exactly in
the digenite field delineated by Yund & Kullerud (1966), but close to the most Fe-rich chalcocite
composition. Traces of Hg, As, Sb, Bi, Se and Te are present, but are mostly below the detection
limit.
14.14.7 Millerite
Millerite is a comparably common accessory phase of the carbonatitic breccia. Representative
analyses are listed in Table A.5.1.33 in the Appendix A.5.1. Average compositions are given in
Table 14.7. Compositional variations of the investigated millerites are illustrated in the ternary
diagram Co-Ni-Fe (at.%; Fig. 14.23).
sample; wt.%
Ku-97-24
N=2
Ku-97-24
Ku-97-24
N=2
Cu
lamellae in sg
matrix, isolated
late-stage veins
Fe
Mn
Co
Ni
As
Sb
S
Se
Total
0.00
0.00
0.00
0.00
0.00
2.60
0.45
0.37
2.87
0.14
0.00
0.00
0.00
0.00
0.00
1.32
0.09
0.23
0.27
0.00
62.30
0.30
63.50
62.40
0.90
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
35.40
0.20
35.20
35.45
0.35
0.00
0.00
0.06
0.03
0.03
101.62
0.46
99.36
101.01
1.38
Ku-98-56
N=2
lamellae in pd
0.08
0.04
1.56
0.15
0.02
0.01
1.49
0.32
60.17
1.04
0.00
0.00
0.00
0.00
36.53
0.31
0.03
0.03
99.92
0.36
Ku-98-130°
N=36
Ku-98-130a
N=27
Ku-98-130a
N=14
matrix, next to cpy
0.08
0.08
0.21
0.22
0.36
0.14
0.27
0.15
0.35
0.23
0.97
0.23
0.01
0.02
0.01
0.01
0.01
0.01
0.15
0.03
0.56
0.03
0.53
0.03
64.18
0.36
63.28
0.55
62.80
0.43
0.01
0.02
0.01
0.02
0.02
0.03
0.01
0.02
0.01
0.02
0.01
0.02
35.60
0.17
35.36
0.26
35.66
0.15
0.11
0.05
0.10
0.05
0.16
0.05
100.42
0.38
99.89
0.59
100.52
0.32
matrix next to py
matrix, next to mag
Table 14.7: Average compositions millerite (σ standard deviation; values for sample Ku-97-24 are taken from von
Seckendorff et al., 2000).
The investigated millerites display significant differences in both their Fe and Co
contents depending on their respective textural positions, i.e.
(i)
as lamellae in ferroan siegenite (Co: 1.2- 1.5 wt.%; Fe: 2.1-3.1 wt.%; Fig. 14.23 b),
(ii)
as a constituent of late sulphide veins (Co: 0.2-0.3 wt.%; Fe: 2.7-3.0 wt.%; Fig.
14.23b),
238
(iii)
(v) as lamellae in polydymite-violarite (Co: 1.2-1.8 wt.%; Fe: 1.4-1.7 wt.%; Fig.
14.23c),
Co
50
50
sg
pd-vio
mi
a
50
Ni
Fe
Co
Co
b
<
<
90
Ku-97-24
matrix
late-stage veins
lamellae in sg
95
95
90
<
Fe Ni
95
Co
<
Ku-98-130a
matrix, next to py
matrix, next to mag
90
Ku-98-130a
matrix,
replaced by cpy
e
d
95
95
95
90
<
Ni
Fe
Co
<
90
90
Fe Ni
95
90
<
Ni
c
<
95
Ku-98-56
lamellae in pd
90
Fe
Fig. 14.23: Compositional variation of millerite from the studied samples, expressed in the ternary diagram Co-NiFe (at.%). a) Dark grey fields designate the compositions of millerites, coexisting with ferroan siegenite in sample
Ku-97-24 (von Seckendorff et al., 2000), light grey fields designate the compositional variation of violaritepolydymite bearing millerite-lamellae in sample Ku-98-56. b) Enlarged section displaying the compositional
variation of millerite in sample Ku-97-24 (von Seckendorff et al., 2000), c) Enlarged section displaying the
compositional variation of millerite in sample Ku-98-56, d) and e) Enlarged section displaying the compositional
variation of millerites in sample Ku-98-130a.
(iv)
as matrix grains bordered by pyrite (Co: 0.5-0.6 wt.%; Fe: 0.1-1.0 wt.%; Fig. 14.23d),
(v)
239
as matrix grains bordered by magnetite (Co: 0.5-0.6 wt.%; Fe: 0.7-1.4 wt.%; Fig.
14.23d),
(vi)
as isolated matrix grains, subsequently replaced by chalcopyrite and/or fletcherite
(Co: 0.1-0.2 wt.%; Fe: 0.1-0.8 wt.%; Fig. 14.23e).
Remarkably, millerites in assemblage (i) and (iii) reveal quite similar Fe:Co ratios like
the coexisting ferroan siegenite and polydymite-violarite, respectively (Fig. 14.23a). The
generally low contents of both Fe and Co of the matrix millerites (iv), (v) and (vi) are in good
agreement with their precipitation from Ni-rich and Co-poor fluids, whereas the elevated Fecontents of millerites in the subsequently formed sulphide veins point to changes in the physicochemical conditions during fenitization, extending the availability of iron in late-stage fluids.
This interpretation is constrained by concentric zoning patterns of matrix millerites in textures
(iv) and (vi), involving a rimward increase of the Fe-contents of millerite towards the bordering
pyrite and chalcopyrite, respectively.
14.14.8 Siegenite
Siegenite was observed in one sample of the carbonatitic breccia (Ku-97-24) by von
Seckendorff & Drüppel (1999) and von Seckendorff et al. (2000). In addition to its occurrence in
the matrix, siegenite may form a constituent of late-stage sulphide veins. In all textural positions,
siegenite contains lamellae of millerite. The formula of siegenite has been calculated on a 4sulphur basis (Table A.5.1.34 in the Appendix A.5.1). EMP analyses of siegenite generally
reveal high Fe contents of 7-12 wt. %. Traces of Cu may reach values of up to 0.5 wt.%.
14.14.9 Violarite-polydymite
In the carbonatitic breccia sample Ku-97-24, polydymite was found as a replacement product of
millerite (von Seckendorff et al., 2000), whereas millerite in sample Ku-98-56 forms lamellae in
polydymite-violarite grains. Formulae of polydymite-violarite have been calculated on a 4
sulphur basis. Representative analyses are listed in Table A.5.1.35 in the Appendix A.5.1.
With Fe-contents of 2.1-3.2 wt.% and Co-contents of 0.2-0.4 wt.% polydymite of sample
Ku-97-24 has a composition near to the polydymite end-member (von Seckendorff et al., 2000).
240
In clear contrast, thiospinels of sample Ku-98-56 display quite variable ratios of Fe/Ni and
Fe/Co. Their formulae can be expressed as Fe0.3Co0.3Ni2.4S4 to Fe0.7Co0.3Ni2.0S4. Most grains are
oscillatory zoned, pointing to episodically changing chemical compositions of the fenitizing
fluids. The compositional range of the thiospinels can be expressed in the subsystem Fe-Co-Ni
forming part of the complex Cu-Fe-Co-Ni system (Fig. 14.24), where the thiospinels of sample
Ku-98-56 predominantly plot within the field of the polydymite-violarite solid solution series
(Ni3S4-Fe3S4).
Cu3S4
Fe3S4
Fe3S4
fletcherite
carrollite
carrollite
fletcherite
Ni-linnaeite
Co3S4
siegenite
Ni3S4
linnaeite
polydymite
violarite
mol.%
CB,so
Ku-97-24
Ku-98-56
Ku-98-130a
literature data
Wagner (1999)
Fe3S4
greigite
Fig. 14.24: Comparison of analysed polydymite-violarite and fletcherite with thiospinel from the literature in the
pseudo-quaternary diagram Co3S4-Ni3S4-Cu3S4-Fe3S4 (mol.%). Literature data (black squares) were compiled by
Wagner (1999) using Craig (1971), Vaughan et al. (1971), Craig & Higgins (1975), Minceva-Stefanova & Kostov
(1976), Minceva-Stefanova (1978), Ostwald (1978), Craig et al. (1979), Riley (1980), D'Orey (1983), Pignolet &
Hagni (1983), Vaughan & Craig (1985), Ixer & Stanley (1996), and references therein. Analyses of thiospinels from
sample Ku-97-24 (von Seckendorff et al., 2000; light grey shaded field) and of carrolites (Wagner, 1999; dark grey
shaded field) are marked for comparison.
241
14.14.10 Fletcherite
Fletcherite occurs in one sample of the carbonatitic breccia (Ku-98-130a), where it replaces
millerite along its grain margins and along cracks. The formula of fletcherite has been calculated
on a 4 sulphur basis (Table A.5.1.36 in the Appendix A.5.1).
1.20
Cu
Fe
Co
1.00
0.80
atoms p.f.u.
0.60
0.40
0.20
0.10
0.08
0.06
0.04
0.02
0.00
2.0
2.2
2.4
2.6
2.8
3.0
3.2
Ni (a.p.f.u.)
Fig. 14.25: Correlation patterns of Cu, Fe and Ni p.f.u. (formula calculation based on 4 sulphurs) in fletcherite from
the REE-rich carbonatitic breccia (Ku-98-130a).
The investigated fletcherites have a quite conspicuous composition, implying the
existence of a new solid-solution series in the Cu-Fe-Co-Ni-S system (Drüppel & Wagner, in
prep.; Fig. 14.24). The fletcherites are near-binary, generally with compositions of Cu0.2Ni3.1S4 –
Cu1.1Ni2.1S4 and low Fe- and Co-contents of 0.0-1.6 wt. % and 0.1-0.7 wt.%, respectively. Most
grains exhibit an asymmetric zoning pattern, involving decreases of Ni and increases of Cu with
increasing distance from millerite. Thus, the typifying feature of the investigated fletcherites is a
pronounced Cu substitution into the structure, reaching a maximum of 21.4 wt.% at the grain
margins, corresponding to 1.1 Cu atoms per formula unit (Fig. 14.25). The solid-solution series
extends from the end-member polydymite towards what is considered as Co- and Fe-poor
242
fletcherite. The substitution scheme for the observed linear correlation trends, relating Ni and Cu
but not Ni and Co or Ni and Fe, suggests a substitution of Cu2+ for Ni2+ on the tetrahedral sites of
thiospinel.
Remarkably, fletcherite has only been observed in sample Ku-98-130a, where millerite is
intimately associated by chalcopyrite, thus suggesting that the replacement of millerite by
fletcherite is accompanied by the breakdown of chalcopyrite, the latter providing the Cu
incorporated by the Cu-Ni fletcherite.
14.15 Barite
Barite has been observed in one sample of the REE-rich carbonatitic breccia (Ku-98-130b),
where the very fine-grained sulphate is localised in the interstices between large ankerite grains
and brownish carbocernaite aggregates. Representative analyses are given in Table A.5.1.37 in
the Appendix A.5.1.
The composition of the analysed barites differs from the ideal formula in revealing
significant amounts of both CaO (0.2-1.4 wt.%) and Ce2O3 (2.5-3.1 wt.%), whereas the Sr
contents are near the detection limit, thus suggesting, that the available Sr was incorporated by
the synchronously crystallising strontianite. The precipitation of barite may be related to latestage oxidation of the reduced sulphur species to sulphate in the evolved carbonatitic solutions
carrying significant amounts of both Ba and Ce. This interpretation is in good accordance with
the high LREE contents of the co-crystallising strontianite. The fact that barite occurs as a
texturally late phase in the REE-carbonatite, whereas all other S-bearing minerals predating the
barite formation are sulphides, strongly suggests an increase of the oxidising conditions during
carbonatite evolution.
14.16 Unspecified Na-Al phase
The unspecified Na-Al phase occurs as a late-stage alteration product of sodalite in the sodaliterich carbonatitic breccia, where it replaces the sodalite along its margins and cracks. The
transparent Na-Al phase is characterized by a fibrous morphology, with the hair-like needles of <
243
0.1 µm in width and 0.5-3 µm in length, commonly found in radiating clusters. Regarding the
above mentioned properties the phase strongly resembles natrolite, which indeed is a common
alteration product of sodalite.
3500
sdl
Intensity (counts)
3000
2500
2000
1500
1000
sdl
500
?
sdl
sdl
?
??
0
sdl
sdl
?
? ? sdl
?
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
3500
Int ensity (counts)
3000
2500
2000
1500
1000
sdl
sdl
sdl
500
sdl
sdl
sdl
sdl
sdl
sdl
sdl
sdl
sdl
sdl
0
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68
2 Theta (°)
Fig. 14.26: Powder X-ray diffractiometry spectra of the unspecified Na-Al phase (peaks marked as ?), which is
intimately intergrown with sodalite (peaks marked as sdl).
Remarkably, however, the Na-Al phase contains large amounts of Al (40-43 wt.%) and
Na (7.6-9.5 wt.%) with a molecular ratio of approximately 4:1 (Table A.5.1.38 in the Appendix
A.5.1), whereas the amounts of all other elements detectable with the help of EMP (EDA
spectra), including Si, are negligible (Table A.5.1.38 in the Appendix A.5.1) and, if analysed, are
obviously caused by the interference with the intimately intergrown sodalite. Thus, the elements
remaining as possible additional constituents of the Na-Al phase include H, C, N and O. Only
two natural occurring phases are yet known, being composed of the above mentioned chemical
species, i.e. dawsonite NaAlCO3(OH)2 and diaoyudaoite NaAl11O17. However, the stoichiometry
244
of both minerals differs significantly from that of the Na-Al phase. The same holds true for yet
known synthetic Na-Al phases. A best-fit formula was reached by calculating the Na-Al phase as
an oxide, leaving a deficiency of 10 wt.% in the total sum of oxides, which was subsequently
converted into a corresponding amount of H2O. Following this calculation, the formula would be
close to NaAl4O6.5ּ1.5H2O.
In order to constrain its crystal structure, the Na-Al phase was separated from the rock
samples Ku-98-18 and Ku-98-130b, powdered and investigated with the help of powder X-ray
diffractiometry (XRD). However, due to the intimate intergrowth of the Na-Al phase with
sodalite, measurements yielded insufficient results, with the spectra being largely dominated by
the sodalite patterns. Thus, only minor and very small peaks deviating from those of sodalite
were identified, which may be representative of the Na-Al phase (Fig. 14.26). Remarkably,
however, none of the yet known synthetic and natural phases fits the remaining 2 Theta values,
thus suggesting that the Na-Al phase represents a new mineral species.
The Si-free composition of the phase and its high Al:Na ratio of 4:1 when compared to
sodalite (Al/Na < 1) strongly suggests that even the late-stage carbonatite fluids were still
strongly deficient in Si but contained only minor amounts of Na.
245
14.17 Trace and rare earth element chemistry of selected minerals – a combined
LA-ICPMS – SRXRF study
14.17.1 Carbonates
In order to investigate the trace element composition of carbonates, formed during different
stages of carbonatite evolution, magmatic calcite from one sample of the carbonatitic breccia
(Ku-98-14) as well as magmatic ankerites from two samples of the carbonatitic breccia (CB,l:
Ku-98-18 and CB,REE: Ku-98-130a) and one ferrocarbonatite sample (Ku-01-04) were analysed
for rare-earth and other trace elements with the help of high-resolution synchrotron micro-XRF.
Analytical results for ankerite and calcite are compiled in Tables A.5.2.1 and A.5.2.2,
respectively, in the Appendix A.5.2.
Interstitial calcite of the carbonatitic breccia sample Ku-98-14 is rich in SrO (1.6-1.8
wt.%) and the LREE (La: 0.5-0.6 wt.%, Ce: 0.9-1.3 wt.%, Nd: 0.4-0.5 wt.%, Sm: 0.03-0.06
wt.%, Eu: 0.02-0.03 wt.%, Gd: 0.02-0.03), whereas Zn (8-31 ppm), Kr (6-11 ppm), Nb (16-54
ppm), Ba (0-196 ppm) and the HREE (Tb: 22-40 ppm, Dy: 151-240 ppm, Ho: 9-27 ppm, Er: 5694, Tm: 0-9 ppm, Yb: 49-78 ppm, Lu-14-39 ppm) are present in only minor amounts. The
analysed calcite displays a concentric zoning pattern, involving a core preference for the REE
and Sr, whereas Ca, Fe and Mn trend to rim. The ΣREE in calcite ranges between 1.9 wt.% (rim)
and 2.5 wt.% (core). The calcite is characterized by positively correlated (La/Nd)
cn
and
(La/Yb)cn ratios increasing from 2.3 (core) to 2.6 (rim) and from 50 (core) to 78 (rim),
respectively. The calcites display straight REE patterns and strong positive (Eu/Eu*)cn anomalies
(1.7-1.9). An accentuated negative trough for Sm is developed for all REE patterns.
Regarding their Fe, Mg and Ca contents all investigated ankerite crystals show a quite
homogeneous composition with the XMg ranging between 0.57 and 0.64. Traces of Zn (6-290
ppm), Kr (8-116 ppm) and Nb (6-132 ppm) are present in all analysed ankerites. However, the
carbonates show considerable variations of the REE between individual samples.
(1) Anhedral matrix ankerite from the layered carbonatitic breccia (Ku-98-14) has a
ΣREE ranging between 0.1 wt.% and 0.8 wt.%, distinctly lower than those of the coexisting
calcite. The REE patterns of the ankerites are straight, with high (La/Nd) cn and (La/Yb)cn ratios
of 2.6-2.9 and 42, respectively (Fig. 14.27a). A strong positive (Eu/Eu*)cn anomaly (1.8) and
negative Sm and Tm anomalies have been observed for one ankerite with elevated REE contents.
246
100000
CB, l; Ku-98-14
calcite
ankerite
sample/chondrite
10000
Cc
Ank
1000
100
10
CB
(N=3)
a)
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy Ho
Er Tm Yb Lu
100000
CB, REE; Ku-98-130b
ankerite
sample/chondrite
10000
CB,REE
(N=2)
1000
Ank
100
10
b)
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy Ho
Er Tm Yb Lu
100000
FV; Ku-01-04
ankerite
sample/chondrite
10000
Ank
1000
100
FV
10
c)
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy Ho
Er Tm Yb Lu
Fig. 14.27: Chondrite-normalized rare-earth element patterns of carbonates from the Swartbooisdrif carbonatitic
breccia (analyzed by SRXRF). Chondrite-normalized rare-earth element patterns of the respective rock samples
(grey; analyzed by ICP-AES and ICPMS) are illustrated for comparison. a) Layered carbonatitic breccia. b) REErich sample of the sodalite-rich carbonatitic breccia. c) Ferrocarbonatite vein.
247
The REE patterns of ankerite are in excellent accordance with those obtained for calcite and the
corresponding rock samples, even though the totals of REE in ankerite are lower than those of
calcite and higher than those of the carbonatitic breccia.
(2) Matrix ankerite from a REE-rich sample of the sodalite-rich carbonatitic breccia
(Ku-98-130b) has comparably low ΣREE of 0.1-0.3 wt.%, which, however, may increase to
values of up to 0.8 wt.% when reaching bordering carbocernaites. Their (La/Nd)
cn
ratios are
generally high but variable, ranging between 1.1-5.4. Values of REE heavier than Nd are below
the detection limit. Regarding the high (La/Nd) cn ratios, the REE patterns of the ankerites agree
well with those of the corresponding rock samples of the REE-rich carbonatitic breccia (Fig.
14.27b). However, in clear contrast to the latter, ankerites contain distinctly lower totals of REE,
thus suggesting that the high REE amounts of the rock samples are rather caused by high modal
amounts of carbocernaite. SRXRF analyses of carbocernaite were performed, but attained mostly
unreasonable and highly variable trace element values.
(3) With values of 0.1-0.2 wt.%, anhedral grains of matrix ankerite from the late
ferrocarbonatite vein (Ku-01-04) reveal the lowest average ΣREE of all ankerites investigated.
Values of REE heavier than Nd are were too low to be detected with help of SRXRF (Fig.
14.27c). The (La/Nd)cn ratios of the ankerites (1.1-1.4) are comparably low, in good accordance
with those of the corresponding rock sample, which, however, contains significantly lower
ΣREE.
In summary, the following evolutionary trend can be established for carbonates of the
Swartbooisdrif carbonatites. Ankerites from the first carbonatite generation, the carbonatitic
breccia, generally reveal higher ΣREE and (La/Nd)cn ratios than those of the late ferrocarbonatite
veins. The LREE contents of ankerite from both the carbonatitic breccia and the ferrocarbonatite
veins are higher than those of the corresponding rock samples, thus suggesting that the
carbonates contain almost the bulk of LREE of the respective samples. In clear contrast,
ankerites from the REE-rich samples of the carbonatitic breccia display distinctly lower ΣREE
and (La/Nd)cn ratios than the corresponding rock samples, hence implying that the high overall
amounts of LREE in these samples are caused by their high modal amounts of carbocernaite.
The fact, that ΣREE of the early crystallized ankerites from the REE-rich carbonatitic breccia are
in the range of those of ankerites from the ferrocarbonatite veins and even lower than those of
ankerites from the REE-poor carbonatitic breccia, strongly suggests that the enrichment of REE
in these rock samples is not a primary feature of the carbonatite melt but was caused by
secondary processes, such as hydrothermal alteration.
248
14.17.2 Sodalite
In order to constrain the origin of sodalite from the sodalite-rich carbonatitic breccia, X-ray
fluorescence (XRF) measurements of hand-picked sodalite separates and in situ LA-ICPMS
analyses have been performed. XRF analyses of sodalite separates, however, attained mostly
unreasonable and highly variable trace element values, resulting from a variable but generally
high amount of mineral and fluid inclusions contained in the sodalite and are thus excluded.
Analytical results for the LA-ICPMS analysis are compiled in Table A.5.3.1 in the Appendix
A.5.3. A description of the analytical technique is given in the Appendix A.2.5
100.0
sample/chondrite
CB,so
sodalite
10.0
anorthosites & leucotroctolites
of the dark anorthosite suite
(N=8)
1.0
0.1
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
Fig. 14.28: Chondrite-normalized rare-earth element patterns of sodalite from the Swartbooisdrif carbonatitic
breccia and anorthosites of the Kunene Intrusive Complex (analysed by LA-ICPMS and ICP-AES, respectively).
The REE contents of sodalite determined by LA-ICPMS are generally low (ΣREE: 2.53.3 ppm), whereas remarkably high values of As (422-657 ppm), Ti (38-57 ppm), Sn (14-51
ppm) and Ba (57-92 ppm) have been recorded. The Rb and Sr contents of sodalite are in the
range of 9-13 ppm and 10-14 ppm, respectively. Regarding their REE-patterns, all analysed
sodalites are characterized by an enrichment of the light REE ((La/Yb)cn: 4.6-12.1) and flattening
in the heavy REE (Fig. 14.28). All REE-patterns display a strong positive Eu anomaly
((Eu/Eu*)cn: 1.5-3.3) although the Ca-contents of sodalite are below the detection limit. These
REE-patterns are quite similar to those obtained for anorthosite and leucotroctolite samples of
the dark anorthosite suite, thus suggesting that most of the sodalite, abundant in the sodalite-rich
249
carbonatitic breccia, was formed at the expense of plagioclase of both the bordering anorthosites
and the incorporated anorthositic xenoliths. Following this argument, the relative decrease of the
absolute amounts of the REE in the sodalite may be related to a volume increase during the
transformation of plagioclase (specific gravity: 2.7 g/cm3) into the less dense sodalite (specific
gravity: 2.3 g/cm3).
14.17.3 Biotite
In order to investigate the rare earth and trace element composition of biotite coexisting with
apatite in samples of the carbonatitic breccia (Ku-98-14), SRXRF analyses were performed
(Table A.5.2.3 in the Appendix A.5.2).
However, the REE contents of all analysed biotites remain below the detection limit, thus
confirming that almost the bulk of the available REE was incorporated by synchronously
crystallising apatite and/or ankerite. Minor amounts of Zn (881-1654 ppm), Rb (282-524 ppm),
Sr (13-72 ppm), Sn (35-287 ppm), Ba (100-5200 ppm) were detected, displaying oscillatory
zoning in the analysed biotite grains. A core preference for Zn was observed for one biotite,
which is inversely correlated with Fe. Ba trends to rim, whereas a patchy to oscillatory zoning
was observed for the Rb, Sr and Sn.
14.17.4 Apatite
Apatite is a geochemical and petrogenetic indicator mineral of carbonatites (e.g. Le Bas et al.,
1992; Rae et al., 1996; Bühn et al., 2001), with its REE chemistry suggested to be linked to the
degree of carbonatite differentiation (Bühn et al., 2001). Therefore, the REE contents of optically
zoned apatites from a sodalite-free sample of the layered carbonatitic breccia (Ku-98-14) have
been investigated in some detail with the help of high-resolution synchrotron micro-XRF
(SRXRF) analysis (Drüppel et al., 2002b; Table A.5.2.4 in the Appendix A.5.2).
In sample Ku-98-14, apatite occurs in two different textural positions, i.e. (1) as large
subhedral matrix grains and (2) as medium- to fine-grained inclusions in biotite, which, at the
same time, are characterized by the highest Na contents. All analysed fluorapatites display a
weak chemical zonation, with the preserved cores being concentrically zoned, involving a
decrease of F, Fe and the REE as well as an increase of Ba towards the core-rim contacts. Small
250
apatite grains, lacking a distinct core zoning, as well as rims of large apatite grains are
characterized by an oscillatory zoning with respect to Cl, Mn, Ba, F, and the REE, whereas Fe
and Y increase towards the outer margins. The apatites involve at least two substitution
mechanisms (Fig. 14.29), i.e. (1) the "belovite scheme" Ca2++Ca2+ → Na++REE3+ (Rønsbo,
1989), which is a quite common substitution mechanism of apatites from carbonatites, and, much
less frequently, (2) the "britholite scheme" Ca2++P5+ → Si4++REE3+ (Hogarth, 1989).
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
9.0
“britholite scheme”
0.35
Si+ΣREE (a.p.f.u.)
Na+ΣREE (a.p.f.u.)
0.40
“belovite scheme”
0.30
0.25
0.20
0.15
0.10
0.05
9.2
9.4
9.6
Ca (a.p.f.u.)
9.8
10.0
0.00
15.2
15.3
15.4
15.5
15.6
Ca+P (a.p.f.u.)
15.7
15.8
Fig. 14.29: Correlation patterns of Ca versus Na+REE and Ca+P versus Na+Si (a.p.f.u.; formula calculation based
on 25 oxygens) in apatite from the carbonatitic breccia (Ku-98-14).
The ΣREE in apatite ranges between 0.9 and 3.1 wt.%, with the highest values obtained
for matrix apatite. The fluorapatites have relatively high La contents of 0.2-0.7 wt.% and high,
positively correlated (La/Nd) cn and (La/Yb)cn ratios increasing from 1.5 to 4.4 and from 33 to
240 (except one value of 488), respectively. All fluorapatites display straight REE patterns and
weakly positive to distinctly negative (Eu/Eu*)cn anomalies of 0.3-1.3 (Fig. 14.30). Accentuated
negative Tb, Er and Yb troughs are mainly displayed by apatite included in biotite, but may also
be prominent in matrix apatite when reaching the contacts to the bordering biotite. This depletion
of the HREE may be related to the early crystallization of pyrochlore in the investigated sample,
which, however, was too small to obtain reliable results with the help of EMP or SRXRF
analysis. With values of 26-163 ppm the Y concentrations of all fluorapatites are extremely low,
with the Y/Ho ratios (0.6-7.8) lying even below the chondritic value of 28. A weak positive
correlation is seen between the Sr and the Y contents of the fluorapatites.
The high ΣREE, (La/Nd) cn and (La/Yb)cn ratios and extremely low Y/Ho ratios of the
analysed apatites as well as their straight REE patterns agree well with the findings of Bühn et al.
(2001) for fluorapatites from strongly fractionated carbonatites. The wide range of (Eu/Eu*)cn
anomalies (0.3-1.3) of the fluorapatites, which are not displayed by the whole-rocks, may
251
indicate the presence of coexisting hydrous fluids which might have extracted Eu2+ from the
carbonatite melt, a process proposed by Bühn et al. (2001). In contrast, the accentuated negative
anomalies for Tb, Er and Yb displayed by the Swartbooisdrif fluorapatites have not been
described so far for other apatites from carbonatites world-wide. The fact, that none of the
anomalies is observed in the corresponding whole rocks may either suggest partitioning of the
elements into a coexisting fluid, or, more probable, fractionation of the compatible elements Tb,
Er and Yb in a HREE-rich phase, such as early crystallising pyrochlore.
100000
100000
matrix apatite
(ap2006)
10000
10000
sample/chondrite
sample/chondrite
matrix apatite
(ap1001)
1000
100
1000
100
b
a
10
10
Y
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
100000
Y
100000
10000
Ap inclusion in biotite
(ap2004)
sample/chondrite
sample/chondrite
(ap2001)
1000
100
10000
1000
100
c
d
10
10
Y
Y
100000
100000
f
(ap2005)
10000
sample/chondrite
sample/chondrite
1000
100
10000
1000
CB
(N=3)
100
e
10
10
Y
La Ce Pr
Y
La Ce Pr
Fig. 14.30: a) to e) Chondrite-normalized rare-earth element patterns of apatite from the Swartbooisdrif carbonatitic
breccia (analysed by SRXRF). f) REE-patterns of the carbonatitic breccia itself (ICP-AES analyses) are given for
comparison.
Figure 14.31 compares the trace-element composition of the Swartbooisdrif carbonatites
with that of fluorapatites from calciocarbonatites and from metasomatized country rocks of other
252
carbonatite complexes elsewhere investigated by Bühn et al. (2001). The (La/Nd)
and
cn
(La/Yb)cn ratios as well as Lacn and the (La/Yb)cn ratios display a positive correlation whereas
CaO and ΣREEcn are negatively correlated, in good accordance with the findings of Bühn et al.
(2001) for magmatic fluorapatites from calciocarbonatites. However, in clear contrast to the
latter, the Swartbooisdrif fluorapatites are characterized by a weak positive correlation of the
(La/Yb)cn and (Eu/Eu*)cn ratios, which was also observed for metasomatically formed
fluorapatites from the Okoruso carbonatite, Namibia (Bühn et al., 2001). This distinct behaviour
of
the
Swartbooisdrif
apatites,
when
compared
to
orthomagmatic
apatites
from
calciocarbonatites, may imply that the fractionation mechanisms in ferrocarbonatite magmas
differ from those of calciocarbonatites with respect to Eu.
6
200000
matrix apatite
150000
apatite inclusion
in biotite
REEcn
(La/Nd)cn
(L
4
100000
2
50000
a
b
0
0
1
10
100
1000
46
(La/Yb)cn
48
50
52
54
56
CaO (wt.%)
2
1000
100
(La/Yb)cn
(Eu/Eu*)cn
1.5
1
10
0.5
d
c
0
1
10
100
1000
1
1000
10000
100000
Lacn
Fig. 14.31: Trace element characteristics of the Swartbooisdrif fluorapatite compared the data of Bühn et al. (2001)
for fluorapatites from calciocarbonatites (light grey shaded fields), from a brecciated fluorapatite layer in
calciocarbonatite (dark grey shaded fields) and from a metasomatized marble country rock in contact to a
calciocarbonatite (middle grey shaded fields).
14.17.5 Pyrochlore
Optically zoned pyrochlore crystals of the sample Ku-01-04, taken from a late-stage
ferrocarbonatite vein, have been investigated in some detail with the help of SRXRF analysis
253
(Table A.5.2.5 in the Appendix A.5.2), which generally reveals high amounts of Cr (0.2-3.2
wt.%), Ta (0.2-0.9 wt.%), W (0.2-0.8 wt.%) and part of the REE (La: 0.4-1.1 wt.%, Ce: 0.8-2.3
100000
FV (Ku-01-04)
pyrochlore 1
sample/chondrite
10000
Pcl
(N=7)
1000
100
FV (Ku-01-04)
10
a)
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy Ho
Er Tm Yb Lu
100000
FV (Ku-01-04)
pyrochlore 2
sample/chondrite
10000
Pcl
(N=9)
1000
100
FV (Ku-01-04)
10
b)
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy Ho
Er Tm Yb Lu
100000
FV (Ku-01-04)
pyrochlore 3
sample/chondrite
10000
1000
FV (Ku-01-04)
Pcl
(N=28)
100
10
c)
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy Ho
Er Tm Yb Lu
Fig. 14.32: Chondrite-normalized rare-earth element patterns of pyrochlore from a ferrocarbonatite vein (Ku-0104), analysed by SRXRF. REE-patterns of the corresponding whole-rock (ICP-AES analyses) are given for
comparison. a) and b) Small matrix grains of pyrochlore (ca. 0.1 mm). c) large pyrochlore grain (0.9 mm).
254
wt.%, Nd: 0.2-0.8 wt.%, Er: 6-201 ppm, Tm: 243-1308 ppm, Yb: 24-375 ppm). The fine-scaled
oscillatory zoning of pyrochlore, observed microscopically, is also displayed by its trace element
chemistry. All analysed grains show a concentric zonation pattern with a rim preference for the
LREE, Yb and Ta, whereas Er trends to core. Even though no reliable results were obtained with
respect to the remaining REE it can be expected that the pyrochlores additionally contain a
certain amount of Tb and may thus indeed be responsible for the negative Tb, Er and Yb
anomalies observed for fluorapatite.
The ΣREE in pyrochlore ranges between 1.4 and 4.1 wt.%, with the highest values
obtained for the smallest pyrochlore grains (Fig. 14.32a, b). All pyrochlores display straight REE
patterns with high (La/Nd) cn and (La/Yb)cn ratios (1.4-4.4 and 12.5-162, respectively) and strong
positive (Tm/Tm*)cn anomalies (2.3-17.4), which are most accentuated in the larger pyrochlore
grains (Fig. 14.32c). The high Tm amounts of the analysed pyrochlores agree well with the
conspicuous negative Tm anomalies observed for both calcite and ankerite, thus suggesting that
most of the Tm available was incorporated into the early crystallising pyrochlore. The
pronounced LREE enrichment in the pyrochlores reflects the REE distribution in the
ferrocarbonatites.
15 Geochemistry
255
15 Geochemistry
15.1 Major and trace element geochemistry
Geochemical data were obtained for 5 rock samples of the fenitized anorthosites, 8 rock samples
of the fenitized syenites, 1 rock sample of the nepheline syenite, 21 rock samples of the
carbonatitic breccia, 3 samples of sodalite accumulations from the carbonatitic breccia and 2
rock samples of the ferrocarbonatite veins. The investigation of the bulk rock geochemistry shall
serve as a tool for constraining the nature and origin of the carbonatitic melts and for
characterizing the carbonatite-induced fenitizing processes, which led to the formation of the
Swartbooisdrif sodalite deposits. Moreover, 2 carbonatites, sampled close to the NamibianAngolan border and presumably belonging to the Mesoproterozoic Lupongola complex of SW
Angola, have been analysed for their major and trace element geochemistry, in order to
investigate a possible genetic relationship between the Swartbooisdrif and the Lupongola
carbonatite centres.
Fenitized anorthosite (A,f)
Ku-98-71, Ku-98-72, Ku-98-77, Ku-98-78, Ku-98-84
Fenitized syenite (S,f)
Ku-98-66, Ku-98-70s, Ku-98-73s, Ku-98-74, Ku-98-103s, Ku-99-09, Ku-99-19s
Fenitized nepheline syenite (S,ne)
Ku-99-15a
Carbonatitic breccia (CB)
Layered carbonatitic breccia (CB,l): Ku-98-05, Ku-98-28, Ku-98-56, Ku-98-70c, Ku-98-118, Ku99-04, Ku-99-17, Ku-99-19c
Massive carbonatitic breccia (CB,m): Ku-98-08, Ku-98-73c, Ku-98-103c
Sodalite-rich carbonatitic breccia (CB,so): Ku-98-17, Ku-98-27, Ku-98-80, Ku-98-83, Ku-98-131
(sodalite-free zone), Ku-99-05, Ku-99-07, Ku-99-08
Sodalite accumulations from the sodalite-rich carbonatitic breccia (Sdl): Ku-98-18, Ku-9847, Ku-01-OD3
REE-rich carbonatitic breccia (CB,REE): Ku-98-130a, Ku-98-130b
256
15 Geochemistry
Ferrocarbonatite veins (FV)
Ku-01-04, Ku-01-05
Carbonatites of the Lupongola complex, NW Namibia – SW Angola
Ku-01-08, Ku-01-10
Geochemical data are listed in Table A.5.4.1 in the Appendix A.5.4. Analytical conditions are
provided in the Appendix A.2.4.
15.1.1 Fenitized anorthosite
The five samples of fenitized anorthosites were taken from two metasomatically overprinted
contact zones between the dark anorthosite suite and large dykes of the carbonatitic breccia. A
principal problem for the interpretation of the geochemical data of the investigated samples is,
that they represent fenitized equivalents of variable anorthosite lithologies (anorthosite and
leucogabbronorite), displaying small-scale compositional variations.
The fenitized anorthosites of both sampled profiles reveal variable SiO2 and Al2O3
amounts of 43-54 wt.% and 20-28 wt.%, respectively, with the lowest values being obtained for
the most altered samples, thus suggesting an interaction between the anorthosites and strongly
Si-deficient fluids (Fig. 15.1). The rocks show enrichment in Fe relative to Mg and strong
increases in LOI (CO2 and H2O), which are reflected by increasing modes of ankerite and
magnetite and the replacement of pyroxene by magnesio-riebeckite and/or biotite. Gains in Fe
are suggested to be caused by the influx of fenitizing solutions carrying considerable amounts of
Fe3+, derived from the carbonatite magma.
Sodium may either increase or decrease with increasing proximity to the carbonatite (Fig.
15.1). It remains uncertain, however, if the apparent depletion of Na in part of the samples (1)
results from the injection of Ca-rich fenitizing fluids or (2) reflects inherited compositional
variations of the magmatic precursors, such as variable ratios of cumulus plagioclase and
intercumulus Fe-Mg silicates. By now possibility (2) is the preferred model, since the
progressive transformation of Ca-rich plagioclase into oligoclase, albite, and sodalite clearly
testifies to the influence of fenitizing solutions carrying considerable amounts of Na.
Calcium decreases with increasing fenitization (Fig. 15.1), even though the modal
amounts of carbonate increase in the same direction. It has to be mentioned, however, that Carich phases of the precursor rocks, such as cumulus plagioclase, intercumulus clinopyroxene and
15 Geochemistry
257
late-magmatic Ca-amphibole, were continuously replaced by sodium-rich minerals during
fenitization, hence suggesting that the solutions were characterized by conspicuously high Na:Ca
ratios.
Profile 1
30
30
wt.%
Ku-98-78
28
26
Profile 2
wt.%
28
Ku-98-72
26
Al2O3
24
24
Ku-98-77
22
Al2O3
22
Ku-98-84
20
20
18
18
Fenitization
12
Ku-98-71
12
wt.%
10
Fenitization
wt.%
CaO
10
FeO
CaO
tot
8
8
6
6
Na2O
MgO
4
4
2
TiO2
LOI
K 2O
0
42
2
0
44 46
48 50 52
SiO 2 (wt.%)
54 56
Na2O
FeOtot
LOI
MgO
TiO 2
K 2O
40 42
44 46
48 50 52
SiO 2 (wt.%)
Fig. 15.1: Compositional changes during the progressive fenitization of anorthositic rocks sampled along two
profiles at the contact zones between anorthosite and the carbonatitic breccia (arrows mark increase of fenitization
of the two sampled profiles).
Trace elements such as V and Zn, incorporated in magnetite, and Co and Ni, incorporated
in pyrite, became strongly enriched during fenitization. Sr is either weakly depleted or strongly
enriched when compared to least altered samples, reflected by variable modes of ankerite and
feldspar. In samples with small amounts of biotite, Ba and K are depleted. Elements which show
an indifferent behaviour during metasomatism are P, Cr and Ga.
258
15 Geochemistry
20.0
4.0
17.5
3.5
fenitized syenite
15.0
3.0
CaO (wt.% )
TiO2 (wt.% )
4.5
A
2.5
2.0
1.5
12.5
A
10.0
7.5
5.0
1.0
0.5
2.5
S
0.0
0.0
25.0
7.5
S
A
MgO (wt.% )
15.0
10.0
5.0
A
2.5
0.0
0.0
7.0
17.5
6.0
15.0
5.0
K2 O (wt.% )
5.0
S
S
FeOtot (wt.% )
Na2O (wt.%)
20.0
4.0
3.0
2.0
A
1.0
S
12.5
A
10.0
7.5
5.0
S
2.5
0.0
0.0
35.0
200
30.0
20.0
Co (ppm)
Al 2 O3 (wt.% )
150
25.0
A
100
15.0
50
S
10.0
S
A
5.0
0
4000
3000
2500
3000
Ba (ppm)
Sr (ppm)
2000
2000
1500
1000
A
500
S
0
30
40
50
S
A
1000
60
SiO 2 (wt.%)
70
0
80
30
40
50
60
70
80
SiO 2 (wt.%)
Fig 15.2: Compositional variation of the fenitized anorthosites and fenitized syenites compared to that of unaltered
samples of the anorthositic suite (dash-dotted grey fields) and the felsic suite (dashed fields), illustrated in selected
Harker-diagrams.
15 Geochemistry
259
When compared to the compositional fields of unaltered samples of the anorthosite suite
(Fig. 15.2) the fenitized anorthosites display a weak tendency towards lower Si, Ca, K and Ba,
whereas Na trends to higher values. However, due to the generally wide compositional scatter of
magmatic
samples
of
the
anorthosite
suite
(anorthosites,
leucotroctolites
and
leucogabbronorites), most of the fenitized anorthosites fall into the fields of their magmatic
counterparts.
The obtained spectrum of chemical compositions, i.e. increases in Na, Mg, Fe, LOI, V,
Zn, Co and Ni accompanied by decreases of Si, Al and Ca, mirrors the continuously changing
modes of the anorthosites during fenitization. Judging from the observed geochemical trends, the
solution responsible for metasomatism of the anorthosites is suggested to be a Si- and Aldeficient Na-rich hydrous fluid, which additionally carried considerable amounts of Fe, Mg, V,
Co, and Ti as well as traces of Ca, K, Sr, Ba, and S, in good accordance with a derivation from a
carbonatite melt.
15.1.2 Fenitized syenite
Fenitized syenites differ from their unaltered counterparts by containing higher average amounts
of Na, Ti, Fe, Ca, P, V, Ni, Sr, Nb, Zr, Th, U, CO2, and H2O and lower contents of Si, K, Rb, and
Ba (Fig. 15.2), reflected by the observed mineralogical changes, i.e. (1) the progressive
replacement of K-feldspar by albite accompanied by minor muscovite (losses: Si, K, Rb, Ba,
H2O; gains: Na), (2) the alteration of anhydrous Fe-Mg silicates to hydrous phases (gains: H2O),
(3) increasing modes of magnetite (gains: Fe, V, Ti), ankerite (gains: Fe, Ca, Sr, CO2), apatite
(gains: Ca, P), pyrite (gains: Fe, S, Ni) and zircon (gains: Zr, Th, U). It remains uncertain,
however, which phase incorporates Nb, since the behaviour of Nb is not linked to any other
element detected with the help of EMP analysis. It may, however, be possible that minor
amounts of pyrochlore were formed during fenitization, which were too small to be observed
with the available analytical facilities. Elements remaining mostly uninfluenced during
fenitization are Al, Mn, Co, Zn, Ga, Y, Pb, La, and Ce. The investigated samples do not display a
clear correlation for the respective elements but a scatter of data points, thus indicating a strongly
varying degree of contamination with Si-deficient and sodium-rich hydrous fluids that were also
capable of transporting appreciable amounts of Fe, Ca, V, Ni, Ti, and Sr as well as traces of Nb,
P, Zr, Th, and U. Even though the proposed composition of this fluid deviates slightly from that
assumed for the solutions which altered the anorthositic samples, both rock suites may have
260
15 Geochemistry
interacted with the same fluid, since the chemical deviations are most likely a function of
compositional differences between the respective source rocks.
Injection of carbonatite magma
24
2.6
wt.%
Al2O3
22
FeO
wt.%
MgO
P2 O5
2.4
tot
MnO
Na 2O
2.2
CaO
20
TiO 2
Al2O3
MgO
2.0
18
1.8
16
1.6
14
1.4
12
1.2
10
1.0
Na2O
8
0.8
MgO
6
0.6
4
FeO tot
2
0.4
0.2
CaO
TiO2
0
50
55
60
SiO 2 (wt.%)
P2O5
MnO
0.0
65
50
55
60
65
SiO 2 (wt.%)
Fig 15.3: Compositional changes of the fenitized syenites, caused by the progressive injection of carbonatite melt
(arrows mark increasing modal amounts of carbonatite minerals in the fenitized syenites).
Samples of the fenitized syenites themselves display distinct geochemical trends (Fig.
15.3), characterized by decreases of Si, accompanied by decreases in Al and Na and increases of
Ca, Fe, Ti, Mn, Sr, and CO3, which are reflected by the enrichment of both ankerite and
magnetite with respect to albite. These trends, however, are not caused by the interaction of the
syenites with fenitizing fluids but presumably result from a progressive injection of
ferrocarbonatite magma, which intermingled with heavily sheared fragments of previously
15 Geochemistry
261
fenitized syenites. It can thus be concluded, that the injection of fenitizing fluids predated the
emplacement of the carbonatite melts for a certain time.
1000
rock / chondrite
100
10
1
0.1
syenite
nepheline syenite
Fig. 15.4: Chondrite-normalized (except Rb, K, P) trace element abundance diagrams (Spider-grams) for the
fenitized syenite (Ku-98-103s) and the fenitized nepheline syenite (Ku-99-15a), using the normalization factors of
Thompson et al. (1983). Compositional ranges of the unaltered samples of the felsic suite (light grey) are marked for
comparison.
Chondrite normalized trace element plots (Fig. 15.4) show depletion in Rb, K, Sr, P, Ti,
and Y. When compared to unaltered samples of the syenite suite, the fenitized syenite sample
contain reveal higher amounts of most of the plotted REE (La, Ce, Sm, and Tb) and distinctly
lower contents of Rb, K, and Y, reflected by the breakdown of K-feldspar during fenitization.
Increases of the REE are in excellent accordance with an interaction between the syenites and
LREE- and Tb-rich, carbonatite-derived fluids.
15.1.3 Fenitized nepheline syenite
The partially fenitized nepheline syenite sample analysed (Ku-99-15a) has an alkali-rich,
peraluminous and silica-undersaturated composition. The ratio FeOtot/(MgO+FeOtot) is 1.0. The
Al, Na, K, and Rb contents are higher than those obtained for the fenitized syenites, whereas the
nepheline syenite exhibits considerably lower amounts of Si, Ca, Fe, Sr, Ba, CO2, H2O, and the
REE.
The chondrite normalized trace element plot of the nepheline syenite (Fig. 15.4) displays
accentuated troughs for Ba, Nb, Ce, Sr, Zr, Ti, and Y and a weak negative slope. The trace-
262
15 Geochemistry
element patterns of the nepheline syenite thus differ significantly from those of all other syenites,
constraining that the two rock types do not belong to one single suite. Remarkably however, the
trace element patterns of the nepheline syenites display similarities (negative anomalies of Ba,
Nb and Ti) with those of mantle-derived melts, weakly contaminated with crustal material
(Thompson et al., 1983). The geochemical characteristics of the nepheline syenite could be
explained by silicate-carbonate liquid immiscibility, even though textural evidence, as listed by
Kjarsgaard & Hamilton (1989), is missing.
15.1.4 Carbonatites
15.1.4.1 Carbonatitic breccia
The 19 samples of the carbonatitic breccia (excluding the REE-rich samples), even though
generally containing significantly lower absolute amounts of SiO2 (22-45 wt.%), mostly display
correlation patterns similar to those of the anorthosites when plotted in selected Harker diagrams
(Fig. 15.5), hence suggesting that anorthosite rather than syenite is the dominant component of
the incorporated fenitized wallrock xenoliths. Only one sample of the layered carbonatitic
breccia (Ku-98-56), containing only minor amounts of carbonatite minerals, reveals relatively
high SiO2 contents (57 wt.%) and plots in the vicinity of the syenite field. The investigated
samples were thus compared to the anorthositic rock suite in order to obtain information about
the geochemical nature of the carbonatite matrix surrounding the anorthositic fragments.
With respect to the anorthositic rocks (Fig. 15.5), the samples of the carbonatitic breccia
(excluding sample Ku-98-56) are strongly depleted in Si, Al and Ca, whereas they contain
distinctly higher contents of Na, reflecting (1) the progressive replacement of plagioclase by
albite and sodalite during fenitization followed by (2) relative decreases of the modal contents of
feldspars and feldspathoids with progressive injection of carbonatite melts. Increasing ratios of
carbonatite matrix/anorthositic xenoliths are mirrored by increases of Fe, Ca, P, S, Sr (up to 3695
ppm), Nb (up to 195 ppm), Zn, Ba, Th, U, the LREE (La: up to 444 ppm, Ce: up to 857 ppm,
Nd: up to 170 ppm), and CO2, thus suggesting that the above mentioned elements represent
primary constituents of the carbonatite magma. High Fe:Ca ratios are taken as evidence that the
carbonatite matrix is broadly ferrocarbonatitic in nature. Sodalite-rich samples of the carbonatitic
breccia (excluding sample Ku-98-131), like the Sdl-separates, trend to high Al and Na, whereas
the sodalite-poor and –free samples of both the layered and the massive carbonatitic breccia as
well as sample Ku-98-131, composed of almost sodalite-free zones from a sodalite-rich
15 Geochemistry
263
20.0
CB,m
CB,m
CB,l
CB,l
CB,so
CB,so
Sdl-poor
Sdl-poor zone
zone
Sdl
4.0
TiO2 (wt.% )
3.5
3.0
2.5
2.0
1.5
A
17.5
15.0
CaO (wt.% )
4.5
12.5
A
10.0
7.5
5.0
1.0
0.5
2.5
S
0.0
0.0
25.0
7.5
S
A
MgO (wt.% )
15.0
10.0
5.0
A
2.5
0.0
0.0
7.0
17.5
6.0
15.0
5.0
K2 O (wt.% )
5.0
S
S
FeOto t (wt.% )
Na2O (wt.%)
20.0
4.0
3.0
2.0
A
1.0
S
12.5
10.0
7.5
A
5.0
S
2.5
0.0
0.0
35.0
200
30.0
Co (ppm)
Al 2 O3 (wt.% )
150
25.0
A
20.0
100
15.0
50
S
10.0
S
A
5.0
3000
4000
2500
3000
Ba (ppm)
Sr (ppm)
2000
2000
1500
1000
A
500
S
0
30
40
50
S
A
1000
60
70
0
80
30
40
SiO 2 (wt.%)
50
60
70
80
SiO 2 (wt.%)
Fig 15.5: Compositional variation of the carbonatitic breccia (excluding samples containing SiO2 < 30 wt.%),
illustrated in selected Harker diagrams and compared to that of sodalite (black dots) as well as to those of unaltered
samples of the anorthositic suite (grey fields) and the felsic suite (white fields).
264
15 Geochemistry
carbonatitic breccia, are generally higher in Ti, Ca, Mg, Fe, and V, caused by lower modal
amounts of sodalite and higher amounts of biotite, magnetite and ankerite.
REE-rich samples of the carbonatitic breccia exhibit low SiO2 (4-11 wt.%) and Al2O3
contents (2-6 wt.%), hence pointing to a minor degree of wallrock contamination (sodalite and/or
albite). Indeed, the Na2O contents (6-8 wt.%) are positively correlated with both Al and Si. The
amounts of FeOtot (4-10 wt.%) and MgO (2-4 wt.%) are too low to classify the investigated
samples as ferrocarbonatites, whereas remarkably high values were obtained for Ca (15-16
wt.%), Sr (3.8-8.9 wt.%), Ba (0.8-2.2 wt.%), La (2.2-5.4 wt.%), Ce (3.3-7.9 wt.%), and Nd (1.13.0 wt.%), correlated negatively with the Si and Al contents and reflected by high modal
amounts of Ca-Sr-REE carbonates. The trace element contents of the REE-rich carbonatite dykes
agree well with the chemical composition obtained for a REE-rich calciocarbonatite dyke in
Rajasthan, India (Wall et al., 1993). However, regarding their major elements, the Swartbooisdrif
dykes are characterized by higher amounts of Al, Fe, Mg, and Na and lower Ca and Mn, due to
the effects of wallrock contamination and their ankeritic nature.
Since the wallrock fragments incorporated in the carbonatitic breccia are almost entirely
composed of albite and/or sodalite (95-100 vol.%), subtraction of the respective oxides (Na2O,
SiO2 and Al2O3) should give approximate values of the composition of the surrounding
carbonatite matrix. Best-fit results were reached by subtracting the totals of Na2O from the
analyses, together with corresponding amounts of SiO2 and Al2O3, calculated for the
petrographically observed albite:sodalite ratios of each single sample (i.e. 0.7-1.0 in the massive
and layered carbonatitic breccia and 0.45-0.55 in the sodalite-rich carbonatitic breccia).
Subsequently, the remaining totals of oxides were recalculated to 100 wt.%, with the determined
normalizing factor being also used for the recalculation of the trace elements incorporated in the
carbonatite matrix. The calculation mode on the basis of Na was chosen, since almost the total of
Na is incorporated in sodalite and albite of the fenitized wallrock fragment, whereas Si and Al
may additionally be present in minerals crystallized from the carbonatite itself, such as biotite.
Samples bearing relict magmatic biotite from the anorthositic suite as well as those containing
abundant muscovite were excluded from the calculation. Recalculated compositions of the
carbonatitic breccia are listed in Table 15.1. Recalculated SiO2 and Al2O3 contents display
ranges of 1.8-9.8 wt.% and 0.5-5.6 wt.%, respectively, with the highest values being obtained for
samples which also contain significant amounts of K2O (0.9-4.5 wt.%), reflected by the
abundance of biotite. Values for FeOtot (13.5-25.0 wt.%), CaO (16.7-28.5 wt.%), MgO (2.9-10.7
wt:%), MnO (0.5-2.1 wt.%), TiO2 (0.6-5.5 wt.%), and P2O5 (0.1-3.0 wt.%) are generally high,
15 Geochemistry
265
but mostly lie within the compositional range of ferrocarbonatites world-wide (see Woolley &
Kempe, 1989, for a review). Sr may reach values of up to 1.2 wt.%.
rock type
sample
CB,l
98-28
CB,l
98-70c
CB,l
99-04
CB,m
98-08
CB,m
CB,m
98-73c 98-103c
CB,so
98-27
CB,so
98-80
CB,so
98-83
CB,so
99-05
CB,so
99-07
CB,so
99-08
wt.%
SiO2
7.82
9.84
4.32
7.45
9.52
7.44
3.64
7.19
1.82
3.42
4.82
4.04
TiO2
4.75
4.31
0.55
5.16
2.94
4.04
5.55
4.61
1.58
0.47
3.86
4.87
Al2O3
8.10
9.32
4.95
7.26
10.73
5.02
2.82
3.77
1.61
3.17
3.13
2.27
Fe2O3
8.48
3.88
5.82
7.41
6.62
6.99
6.39
7.44
3.89
8.15
10.16
5.64
FeO
MnO
MgO
CaO
11.59
1.05
5.82
21.98
17.45
0.82
6.37
21.86
12.29
1.01
10.61
21.62
16.89
0.55
10.18
20.12
13.81
1.54
5.89
21.84
17.67
0.68
9.15
19.32
18.02
1.06
8.00
20.09
18.27
0.85
10.69
16.72
9.94
1.42
7.54
27.98
11.90
2.05
6.39
23.30
11.90
0.96
3.59
25.30
13.66
0.53
2.91
28.52
Na2O
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
K2O
2.39
2.34
2.14
1.53
3.58
4.04
2.39
4.53
0.93
1.62
3.28
3.29
P2O5
0.13
0.46
0.20
0.55
0.07
1.24
0.18
1.74
3.00
0.18
0.62
0.60
S
LOI
b.d.l.
27.89
0.13
23.22
b.d.l.
36.49
0.09
22.82
b.d.l.
23.46
b.d.l.
24.42
0.12
31.75
b.d.l.
24.20
0.10
40.20
b.d.l.
39.36
b.d.l.
32.39
0.60
33.07
CO2
23.67
18.10
28.52
17.82
18.26
21.22
27.09
21.82
35.14
32.16
26.70
26.85
H2O
4.22
5.13
7.97
5.00
5.20
3.20
4.66
2.38
5.05
7.21
5.69
6.21
Sum
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00 100.00
100.00
ppm
V
Cr
Zn
Rb
Sr
Nb
Ba
Pb
Th
U
Ce
Nd
Sum
320
148
211
60
4078
60
722
24
b.d.l.
13
125
b.d.l.
5761
311
120
435
50
2821
135
1399
50
b.d.l.
b.d.l.
505
147
5974
124
43
283
43
4905
159
864
95
b.d.l.
b.d.l.
231
133
6881
511
180
434
43
2301
155
1804
94
137
b.d.l.
548
183
6391
65
32
527
45
4634
234
6343
30
b.d.l.
b.d.l.
b.d.l.
b.d.l.
11909
402
107
596
100
2180
209
2458
80
b.d.l.
b.d.l.
462
243
6838
363
148
463
86
5603
148
888
109
b.d.l.
41
142
65
8055
100.00 100.00
325
356
575
159
3840
387
1259
43
b.d.l.
19
467
194
7623
87
b.d.l.
232
35
11911
461
1431
235
26
19
2763
812
18013
57
43
190
32
6467
312
1591
104
b.d.l.
25
700
233
9756
277
465
362
103
6311
198
1173
106
b.d.l.
b.d.l.
304
85
9384
487
559
374
79
4000
212
1907
140
60
b.d.l.
933
325
9076
Table 15.1: Recalculated composition of the carbonatite matrix of the carbonatitic breccia (see text for details).
Following the IUGS classification (Le Maitre, 1989), which is adopted for this study, the
recalculated samples of the carbonatitic breccia can be classified as ferrocarbonatite (Fig. 15.6a),
but they cross the ferrocarbonatite-ferruginous calciocarbonatite–magnesiocarbonatite borders
after the revised carbonatite nomenclature of Gittins & Harmer (1997; Fig. 15.6b). Similar trends
can be observed for the REE-rich samples of the carbonatitic breccia (Fig. 15.7), even though the
classification of these samples in the ternary diagram Ca-Mg-(Fe+Mn) bears uncertainties, due to
their comparably low amounts of Ca, Mg, Mn, and Fe and the high amounts of REE.
266
15 Geochemistry
(a)
wt.%
(b)
CaO
mol.%
CaO
(sodalite mining area)
calciocarbonatite
calciocarbonatite
REE-rich carbonatitic breccia
ferruginous calciocarbonatite
magnesiocarbonatite
MgO
ferrocarbonatite
FeO* + MnO
magnesiocarbonatite
MgO
ferrocarbonatite
FeO* + MnO
Fig 15.6: Classification diagrams for carbonatites, using their major element composition. a) IUGS classification
(Le Maitre, 1989), b) classification after Gittins & Harmer (1997).
Chondrite normalized trace element plots of the investigated REE-poor samples of the
carbonatitic breccia are illustrated in Fig. 15.7a. Most samples exhibit smooth patterns, peaking
at Nb, La and Ce and displaying variably accentuated troughs at Rb, K, Sr, P, Zr, and Ti. The
negative anomalies point to a K-, Sr-, P-, and Ti-depleted parental melt and thus suggest
extensive biotite, apatite and Fe-Ti oxide crystallization prior to the emplacement of the
carbonatites.
106
10000
REE-rich
104
rock / chondrite
rock / chondrite
1000
100
10
1
1
0.1
102
a)
Sdl-rich
0.01
b)
Fig. 15.7: Chondrite-normalized (except Rb, K, P) trace element abundance diagrams (Spider-grams) for the
carbonatitic breccia. a) REE-poor samples (Ku-98-08, -103c, -118, -131). b) REE-rich samples (Ku-98-130a, Ku98-130b).
Similar patterns are displayed by the REE-rich samples of the carbonatitic breccia (Fig. 15.7b),
even though they contain higher absolute amounts of Ba, Th, La, Ce, Nd, Sm, and Tb, hence
emphasising the negative anomalies at Rb, K, P, and Ti. It may thus be concluded, that the
15 Geochemistry
267
formation of the REE-rich carbonatitic breccia post-dated the continuous crystallization of
ankerite, apatite, biotite and Fe-Ti oxides from the carbonatitic breccia. Following this, the Ba-,
Th- and REE-enrichment of these rocks may have been caused by strong fractionation of the
previously intruded carbonatite melts of the carbonatitic breccia, even though secondary REEenrichment by late-stage, REE-rich carbonatite fluids (autometasomatism) can not be ruled out.
In general, the trace element patterns of all samples of the carbonatitic breccia display
similarities with those characteristic of oceanic nepheline melilitites and basanites (e.g.
Thompson et al., 1983).
15.1.4.2 Ferrocarbonatite veins
Compared to all other rock-types, the late ferrocarbonatite veins (Ku-01-04, -05) are
characterized by the lowest SiO2- (0-1.4 wt.%), Al2O3- (0.2-0.4 wt.%) and Na2O-contents (0-0.2
wt.%) due to the low potential of these small-scaled dykes and veins to incorporate wallrock
material.
(a)
wt.%
(b)
CaO
mol.%
CaO
calciocarbonatite
calciocarbonatite
ferruginous calciocarbonatite
magnesiocarbonatite
MgO
ferrocarbonatite
FeO* + MnO
magnesiocarbonatite
MgO
ferrocarbonatite
FeO* + MnO
Fig 15.8: Classification diagrams for carbonatites, using their major element composition. a) IUGS classification
(Le Maitre, 1989), b) classification after Gittins & Harmer (1997).
Values of Fe, Mg and Ca are similar to the mean values of ferrocarbonatites world-wide,
as summarized by Woolley & Kempe (1989). The MnO contents are generally high (2.33-2.43
wt.%), whereas the amounts of K, Ti, P, and S are near the detection limit, presumably reflecting
fractionation of K-, Ti-, P- and S-bearing minerals like biotite, Fe-Ti oxides, apatite and
268
15 Geochemistry
sulphides from the carbonatitic breccia. The late ferrocarbonatite veins exhibit Sr, La, Ce, and
Nd contents in the ppm range, although Sr may reach up to 0.5 wt.%. According to the
classification of both Le Maitre (1989) and Gittins & Harmer (1997) the rocks are
ferrocarbonatites (Fig. 15.8).
Chondrite normalized trace element plots of two ferrocarbonatite samples reveal quite
different patterns (Fig. 15.9), i.e. (1) a smooth pattern, displaying accentuated troughs at Ba, P
and Ti (sample Ku-01-04), similar to those obtained for the carbonatitic breccia, and (2) a
straight pattern, characterized by a strong positive Sr anomaly and a negative Ti anomaly
(sample Ku-01-05). The observed deviations of the trace-element patterns may reflect the
abundance of pyrochlore in sample Ku-01-04, which is lacking in sample Ku-01-05.
106
Ferrocarbonatite veins
Ku-01-04
Ku-01-05
rock / chondrite
104
2
10
1
0.01
Fig 15.9: Chondrite-normalized (except Rb, K, P) trace element abundance diagrams (spider-grams) in the late
ferrocarbonatite veins.
15.2 Rare earth element geochemistry
REE-analyses were carried out on one sample of the fenitized syenite (Ku-98-103s), one sample
of the fenitized nepheline syenite (Ku-99-15a), four samples of the REE-poor carbonatitic
breccia (Ku-98-08, -103c, -118, -131), two samples of the REE-rich carbonatitic breccia (Ku-98130a, -130b) and two samples of the ferrocarbonatite veins (Ku-01-04, -05). Analytical results
are given in Table A.5.4.2 in the Appendix A.5.4.
15 Geochemistry
269
The fenitized syenite has a ΣREE of 868 ppm and high (La/Nd)cn and (La/Yb)cn ratios of
3.4 and 30.6, respectively. Like its unaltered counterparts, it displays a straight REE patterns
lacking accentuated positive or negative anomalies (Fig. 15.9). When compared to syenites,
which were not affected by metasomatism, the fenitized syenite has higher contents of LREE
(La, Ce, Pr, Nd) whereas the contents of all other REE are almost identical to those of the
unaltered syenites. It can thus be concluded, that the fenitizing fluids were exceptionally capable
of transporting the LREE with respect to the HREE.
With a value of 43 ppm, the ΣREE of the fenitized nepheline syenite is extremely low.
The sample is characterized by a straight REE pattern and comparably low (La/Nd)cn and
(La/Yb)cn ratios of 1.2 and 3.4, respectively (Fig. 15.10). A distinctly negative (Eu/Eu*)cn
anomaly (0.5) suggests plagioclase fractionation prior to the nepheline syenite emplacement. The
REE pattern of the nepheline syenite differs significantly from those of the syenites, thus
constraining that the nepheline syenite represents an independent magmatic unit, which may be
genetically related to the carbonatite suite.
1000000
syenite
nepheline syenite
100000
rock / chondrite
10000
1000
100
10
1
0.1
La
Ce
Pr
Nd Sm Eu
Gd Tb Dy
Ho
Er Tm
Yb Lu
Fig. 15.10: Diagrams showing the variations in chondrite-normalized REE abundances for the fenitized syenite and
the fenitized nepheline syenites compared to those of unaltered samples of the felsic suite (shaded grey).
The ΣREE of the carbonatitic breccia and the ferrocarbonatite veins range between 9
ppm and 17.7 wt.%, with the highest values obtained for the REE-rich carbonatitic breccia and
the lowest for the ferrocarbonatite veins. All investigated samples have relatively high and
270
15 Geochemistry
positively correlated (La/Nd) cn and (La/Yb)cn ratios, increasing from 2 to 6 and from 11 to 454
(except samples of the REE-rich carbonatitic breccia with Yb below the detection limit),
respectively. Both, the carbonatitic breccia and the ferrocarbonatite veins display straight REE
patterns and weak positive (Nd/Nd*)cn anomalies of 1.1-1.6 (Fig. 15.11). The samples of the
REE-rich carbonatitic breccia have Tm-contents similar to those of the REE-poor samples of the
carbonatitic breccia, but substantially more LREE, reflected by the high modal abundance of CaSr-REE carbonates. The ferrocarbonatite sample Ku-01-04 has REE abundances similar to the
carbonatitic breccia, whereas the ferrocarbonatite sample Ku-01-05 reveals significantly lower
REE amounts but a similar overall pattern. The high ΣREE, (La/Nd) cn and (La/Yb)cn ratios and
the straight REE patterns of the REE-rich carbonatitic breccia, when compared to the REE-poor
samples of the carbonatitic breccia and the ferrocarbonatite veins, suggest that they either
represent highly fractionated carbonatite melts or that autometasomatism by late-stage
carbonatite fluids is responsible for the strong enrichment of the LREE. In contrast, the REE
abundances of the carbonatitic breccia and the ferrocarbonatite veins are conspicuously low
when compared to other carbonatites world-wide (e.g. Horning-Kjarsgaard, 1998). With values
of 1-172 ppm the Y concentrations of most samples are extremely low, with the Y/Ho ratios (2029) lying at or even below the chondritic value of 28.
1000000
REE-poor carbonatitic breccia
100000
rock / chondrite
10000
1000
100
10
1
0.1
La
Ce
Pr
Nd Sm Eu Gd Tb Dy Ho
Er Tm Yb Lu
Fig. 15.11: Diagrams showing the variations in chondrite-normalized REE abundances for REE-poor and REE-rich
samples of the carbonatitic breccia and the younger, almost silicate-free ferrocarbonatite veins.
15 Geochemistry
271
15.3 Discussion and interpretation of the carbonatite evolution
To summarize, the composition of the ferrocarbonatite veins is quite similar to that obtained for
the carbonatites forming part of the carbonatitic breccia, even though, in contrast to the latter, the
younger ferrocarbonatites contain lower amounts of Fe, Mg, K, Ti, P and S, which were
presumably incorporated by previously crystallized biotite, ankerite, Fe-Ti oxides, apatite and
sulphides from the carbonatitic breccia. Following this, the ferrocarbonatite veins may be
interpreted as fractionation products of the ferrocarbonatite magma which forms the carbonatitic
breccia. This interpretation is constrained by the (FeOtot+MnO+MgO) vs. DI plot of Fig 15.12a.
The DI ratio of the investigated carbonatites is in general negatively correlated with the
(FeOtot+MnO+MgO), with the highest DI being displayed by the late ferrocarbonatite veins,
which may thus be interpreted to be the most fractionated members of this group. In contrast,
REE-rich samples of the carbonatitic breccia plot in range of the corresponding REE-poor
samples, thus suggesting that the REE-enrichment of the former is not the result of extensive
fractionation, but rather caused by secondary processes.
tot
1000
a
35
b
30
100
25
20
15
ld
fie
10
10
REE-poor carbonatitic breccia
5
0
tite
na
bo
r
ca
La/Yb
FeO +MnO+MgO (wt.% )
40
0.2
0.3
0.4
0.5
0.6
DI
0.7
0.8
0.9
1
1
10
100
La (ppm)
1000
10000
Fig 15.12: Chemical characteristics of the Swartbooisdrif carbonatite samples. a) Plot of DI vs. FeOtot+MnO+MgO
(wt.%) (DI = CaO/(CaO+FeOtot+MnO+MgO) molar ratios; arrow marks fractionation trend; values for the
carbonatitic breccia are derived from recalculated analyses). b) Plot of La vs. La/Yb ratios (carbonatite field after
Andersen, 1987).
Judging from the geochemical data the investigated carbonatite samples are supposed to
represent the crystallization products of ferrocarbonatite magmas. Conformably, in the La vs.
La/Yb diagram (Fig. 15.12b) most samples (except sample Ku-01-05) plot within the carbonatite
field of Andersen (1987).
272
15 Geochemistry
15.4 Comparison with the Lupongola carbonatites
Regarding their major and trace element geochemistry, the Swartbooisdrif carbonatites differ
significantly from the carbonatites sampled from two north-west trending dykes, situated close to
the Namibian-Angolan border ca 2.5 km south-west of the Lupongola Complex, SW Angola.
Sample Ku-01-08, a magnesio-carbonatite, contains traces of Si, Fe, V, Sr, and La and high
amounts of MgO (19.9 wt.%) and CaO (30.9 wt.%), reflecting the modal dominance of dolomite.
All other elements are below the detection limit. In contrast, EMP analyses of sample Ku-01-10,
a calciocarbonatite, reveals minor amounts of SiO2 (1.6 wt.%), Al2O3 (0.5 wt.%) and P2O5 (0.6
wt.%), traces of Zn (70 ppm), Sr (1510 ppm), Y (191 ppm), Nb (201 ppm), Ba (120 ppm), La
(290 ppm), Ce (780 ppm), and Nd (257 ppm), whereas remarkably high values are obtained for
FeO (8.4 wt.%), MnO (2.5 wt.%) and CaO (45.9 wt.%), mirrored by the mineral assemblage of
calcite, magnetite, together with minor apatite. In clear contrast to the Swartbooisdrif
carbonatites, no metasomatic aureoles have been observed in the anorthosites bordering the
magnesio- and calciocarbonatite dykes. If the Lupongola carbonatites indeed belong to the same
magmatic suite like the Swartbooisdrif carbonatites, as has been suggested by several authors
(e.g. Menge, 1998; Thompson et al., 2002), they represent a less fractionated carbonatite unit,
presumably formed previous to the emplacement of the Swartbooisdrif dykes.
16 Mass balance calculations
273
16 Mass Balance Calculations
Solidified carbonatites are the result of complex chains of interrelated processes, such as
differentiation, fractional crystallization, element exchanges between the carbonatite magmas
and coexisting fluids as well as subsolidus re-equilibration (e.g. Le Bas, 1977; Gittins, 1979;
Andersen, 1984, 1987, 1989; Treiman & Essene, 1985; Dawson et al., 1987). Since fenites are
the end-products of metasomatism, compositionally changed by the interaction between a
precursor rock and emanations from carbonatite or alkaline silicate magmas, the investigation of
alteration processes can convey important information about the nature and evolution of both the
carbonatite intrusives and their wall-rocks. However, apparent element gains or losses, observed
in the metasomatic aureoles must not per se infer element mobility but may also be attributable
to volume changes, since elements that behave immobile during hydrothermal alteration are
concentrated during net mass loss and diluted during net mass gains. An accurate calculation of
the material changes during metasomatism therefore requires the knowledge of both, the
compositional and the volume changes during fenitization.
16.1 The Gresens approach
An equation, assessing the compositional and volume changes of a parent rock during
metasomatism has been formulated by Gresens (1967):


 S B 

 − X A 
X n = w ⋅  FV ⋅ X B ⋅ 
 S A 




where:
Xn =
change in weight proportion of component n
w =
initial weight of parent rock (generally set to 100 g, so that
if XA and XB are in percent then Xn is in weight percent
change of the component n)
volume factor (e.g. the volume ratio of altered rock to
parent rock)
weight proportion of component n in parent rock A and
altered rock B, taken from the respective chemical analyses
specific gravities of parent and altered rock, measured on
representative rock chips
FV =
XA, XB =
SA, SB =
The equation contains two unknowns, Xn and Fv. However, a likely value of the volume
factor can be deduced from composition-volume diagrams, with Fv being determined by the
intersection points of lines of elements assumed to be immobile with the zero mass change axis
274
(i.e. Xn = 0; no loss or gain of the respective element). If the values for Fv for several elements
and oxides coincide or only slightly differ, confidence is gained regarding their immobility. The
average volume factor (calculated as an average of the volume factors of oxides and elements
proven to be immobile) may then be applied to the mobile elements to obtain real mass changes.
An alternative method for the determination of Fv with the help of composition-volume
diagrams was proposed by Kresten (1988). The fundamental assumption of Kresten (1988) is,
that the sums of components supplied and removed during metasomatism roughly balance, thus
resulting in minimum total mass changes of a rock during fenitization. Hence, values of minor
and zero mass changes of a rock can be used to determine the corresponding Fv.
16.2 Mass change calculations for the fenitization of anorthosites and syenites
Mass balance calculations have been performed for three samples of the fenitized
syenites and one fenitized anorthosite sample. In addition, six samples of the carbonatitic
breccia, being almost entirely composed of fragments of fenitized wallrock anorthosite, have
been investigated in order to obtain information about the evolution of the fenitizing fluids
during continuous fluid-rock interaction. Composition-volume diagrams of the fenitized
wallrock lithologies and fenite xenoliths show no preferential clustering of intersection points of
the element lines with the zero mass change axis. Thus, values of minimum total mass changes
have been used to determine FV. A good fit was reached for minimum changes of oxygen, which
indeed is generally thought to remain relatively constant during metasomatism. Results of the
mass change calculations for the fenitization of anorthosites and syenites are listed in Tables 16.1
and 16.2, respectively.
The volume factor of 0.99, calculated for the fenitized anorthosite (Ku-98-71), infers a
weak loss in volume of 1 % during fenitization. This volume decrease is most probably created
by increasing modal contents of the high-density minerals magnetite and ankerite, leading to a
density increase from anorthosite (ρ = 2.76) towards the fenitized anorthosite (ρ = 2.81). In clear
contrast, volume factors determined for the transformation of anorthosite into fenitized xenoliths
of the carbonatitic breccia (ρ = 2.56-2.76) mostly imply volume gains of 0.3-5.2 % (Fv: 1.001.085), which increase (1) with decreasing tectonic deformation of the samples and (2) with
increasing modal amounts of sodalite (ρ = 2.27-2.33), formed at the expense of plagioclase (ρ =
2.62-2.76). Conformably, the sodalite-free and heavily tectonised sample Ku-98-103c (ρ = 2.96)
is the only sample of the carbonatitic breccia that reveals a volume loss of 4 % (Fv: 0.96) during
fenitization. Volume factors of 1.01-1.05 were applied to reveal the gains or losses of elements
275
caused by the transformation of syenites (ρ = 2.65-2.77) into fenitized syenites (ρ = 2.67-2.70).
The factors infer gains in volume during fenitization of 1-5 %, which are reflected by increases
of the modal abundance of feldspar with respect to the Fe-Mg silicates. Results of the mass
balance calculations of fenitized anorthosites and fenitized syenites are shown in Fig. 16.1 and
Fig. 16.2, respectively, in order to illustrate the differing activity gradients during the two types
of equilibration.
Rock type
Sample
FV
A / A, f
(98-72 / 98-71)
0.989
changes in g / 100 g
Si
Ti
Al
Fe3+
Fe2+
Fetot
Mn
Mg
Ca
Na
K
P
C
H
O
A / CB
A / CB
A / CB
A / CB
A / CB
A / CB A / mean CB
(98-72/99-04) (98-72/99-05) (98-72/99-07) (98-72/98-08) (98-72/98-73c) (98-72/98-103c)
N(1:6)
1.027
1.031
1.052
1.085
1.003
0.957
1.024
-2.95
0.76
-2.61
1.45
1.82
3.27
0.02
0.03
-2.25
3.04
0.04
0.01
1.04
0.13
0.00
-4.61
-0.01
-3.58
0.92
0.87
1.79
0.22
0.10
-2.51
4.07
0.30
0.02
1.96
0.25
0.00
-3.78
-0.04
-2.69
1.10
0.17
1.26
0.39
-1.01
-3.19
6.10
0.06
0.01
1.72
0.16
0.00
-3.68
0.56
-2.72
1.63
0.37
2.00
0.17
-1.43
-2.41
6.13
0.50
0.07
1.47
0.13
0.00
-4.02
1.25
-4.57
1.83
3.46
5.29
0.14
0.66
-1.37
1.37
0.26
0.09
1.47
0.19
0.00
-3.30
0.59
-2.84
1.40
1.96
3.37
0.44
-0.63
-1.46
1.31
0.89
0.00
1.33
0.18
0.00
-4.13
0.90
-4.46
1.59
3.42
5.01
0.18
0.26
-1.94
2.95
1.11
0.22
1.75
0.09
0.00
-3.92
0.54
-3.47
1.41
1.69
3.10
0.26
-0.35
-2.15
3.67
0.51
0.07
1.62
0.17
0.00
14.88
-14.39
0.48
18.13
-20.15
-2.02
18.29
-19.39
-1.09
18.93
-18.75
0.18
19.62
-19.16
0.46
14.30
-15.50
-1.20
20.58
-19.97
0.61
17.94
-18.53
-0.59
changes in ppm / 100 g
V
151.95
Co
8.23
Ni
5.40
Zn
36.35
Ga
5.11
Rb
Sr
102.20
Y
Zr
34.18
Nb
Ba
241.31
Pb
-6.00
Ce
Nd
-
15.17
-0.68
26.11
67.11
-0.31
14.71
1163.19
15.69
56.88
53.94
98.22
26.36
78.45
45.11
-11.26
-13.36
-59.18
23.13
11.54
8.85
1270.38
43.28
85.57
240.71
22.52
191.79
63.93
54.80
-11.77
-53.84
78.05
6.22
30.30
1363.24
26.26
43.42
58.57
151.39
25.31
89.88
25.25
200.06
18.72
38.46
163.60
8.33
19.26
520.77
29.40
181.44
68.93
605.79
35.56
243.28
81.09
-0.91
-15.93
183.75
1.06
18.06
1368.57
58.20
94.33
2363.99
6.04
-
142.14
4.98
-15.67
222.15
13.73
42.03
417.51
55.36
149.67
88.16
840.37
27.83
194.77
102.51
65.73
-3.09
-12.86
122.04
6.72
22.05
1021.24
25.06
88.22
74.85
713.06
23.91
159.19
63.36
Gains (F.I.)
Losses
G+L
Table 16.1: Calculated values of absolute compositional changes during the fenitization of the anorthosites (A/FA
anorthosite/fenitized anorthosite, A/CB anorthosite/carbonatitic breccia, F.I. fenitization index, Fv assumed volume
factor, G+L sum of gains and losses).
An immediate impression of Table 16.1 and Fig. 16.1 is, that the fenitization of the
wallrock anorthosite was less severe than that of the anorthositic xenoliths forming part of the
carbonatitic breccia. This interpretation is supported by the lesser petrographic and geochemical
276
grades of alteration observed for the fenitized wallrock anorthosites. Remarkably, however, both
fenite types display almost similar chemical trends, suggesting that the composition of the
fenitizing fluids did not change significantly during fluid-rock interaction and hence pointing to
conspicuously high fluid:rock ratios. The main components subtracted during the fenitization of
the anorthosites are Si, Al and Ca, whereas Fe, Na, H2O, CO2, Zn, Sr, and Ba and, to a lesser
content, K, P, V, Zr Ce, and Nd, are the main components added. The ratio of FeOtot/MgO is
relatively high; MgO may be added or, more commonly, subtracted. The fact, that Fe3+ is not
enriched with respect to Fe2+ provides evidence against elevated oxygen fugacities of the
fenitizing fluid.
Sample
Rock type
FV
changes in g / 100 g
Si
Ti
Al
Fe3+
Fe2+
Fetot
Mn
Mg
Ca
Na
K
P
C
H
O
Gains (F.I.)
Losses
G+L
changes in ppm / 100 g
V
Co
Ni
Zn
Ga
Rb
Sr
Y
Zr
Nb
Ba
Pb
Ce
Nd
S / S, f
98-40 / 98-73s
1.024
S / S, f
98-40 / 98-103s
1.047
S / S, f mean S / mean S, f
98-40 / 99-19s
N (5 / 3)
1.010
1.024
-0.89
-0.19
1.38
-1.91
-2.02
-3.93
0.09
0.01
1.10
2.66
-3.19
-0.04
0.88
0.05
2.13
-0.05
-0.10
1.30
-1.06
-1.40
-2.46
-0.07
0.06
0.09
3.44
-3.56
-0.02
0.55
0.07
0.00
-0.09
-0.07
1.31
-0.86
-1.22
-2.07
0.00
0.17
-0.77
2.09
-2.08
0.02
0.10
0.08
-1.76
-0.42
-0.12
1.30
-1.28
-1.55
-2.83
0.01
0.08
0.13
2.71
-2.94
-0.01
0.51
0.06
0.00
7.69
-11.48
-3.79
5.67
-8.01
-2.34
6.68
-6.69
0.00
8.88
-8.48
0.40
-20.88
-56.14
-14.12
750.51
-679.34
379.92
315.03
0.48
-
27.48
7.32
2.17
-77.25
-9.61
459.35
-5.41
1352.17
-53.41
635.39
-0.14
192.64
38.02
12.39
-34.08
-0.02
199.40
-9.80
-27.26
86.04
-5.80
360.95
-47.56
-219.24
0.32
-56.10
-30.03
19.67
-16.16
1.03
22.62
-11.25
-26.68
429.48
-5.85
335.64
92.97
238.22
0.16
65.56
3.14
Table 16.2: Calculated values of absolute compositional changes during the fenitization of the syenites (S/FS
syenite/fenitized syenite, F.I. fenitization index, Fv assumed volume factor, G+L sum of gains and losses; mean S
calculated for samples Ku-98-40, Ku-99-12, -13, -14, -20).
277
As shown in Fig. 16.2, fenitization of the syenites appears to be less pervasive than that
of the anorthosites. Moreover, the major elements of the fenitized syenites show a quite distinct
mobilisation behaviour compared to those of the fenitized anorthosites. Si, Fetot and K were
strongly depleted during fenitization, whereas Al, Na, H2O, and CO2 are the main components
added. Ti, Mn, Mg and P seem to have been constant during metasomatism and Ca displays no
clear mobilisation trends. Regarding the behaviour of the trace elements, the fenitization of the
syenites strongly resembles that of the anorthosites, as Zn, Sr, Ba, and Ce show a strong
enrichment.
Fig. 16.1: Diagrams illustrating the main chemical changes of anorthositic protoliths caused by equilibration with
fenitizing fluids emanating from the carbonatite magma (Y axis: results of the mass balance calculations compiled in
Table 16.1).
However, the observed differences between the element mobilisation behaviour of the
syenites and the anorthosites during alteration must not, per se, indicate compositional
differences between the fluids responsible, but may be better explained by differences in the
activity gradients between the carbonatite magma and the two rock suites, caused by contrasting
bulk-rock compositions of the anorthositic and syenitic protoliths. In general, large
compositional differences between the protolith and the source of fenitization will produce high
gradients of activity, thus enhancing element transfer to the low activity environment until
equilibrium is reached. In case of the anorthosites, the high activity of Ca, Al and Si with respect
to the fenitizing fluid generates an activity gradient that causes the removal of these elements. In
contrast, the low activities of Al and Ca in the syenites may have created an opposite trend,
leading to addition of these components. Following this, the nature of the source of fenitization
may be well constrained, if the fenitization patterns of both the syenites and the anorthosites are
taken into consideration. As is evident from Fig. 16.1 and 16.2, metasomatism of both rock units
provoked increases in Na, CO2, H2O, V, Zn, Sr, Ba Nb, and LREE, irrespective of the
278
compositions of the protoliths, thus suggesting that these elements form major constituents of the
fenitizing fluid. In contrast, both the syenites and the anorthosites suffered a variably severe
desilication, pointing to a very low silica activity of the magma responsible. Elements which,
depending on the protolith composition, may be added or subtracted are Ti, Al, Fe, Mg, Ca, and
K. These elements are thought to have been contained in the fenitizing fluids in only minor
concentrations, thus creating strongly differing activity gradients between the fluid and the
anorthosites and syenites, respectively. The abundance of hydrous ferro-magnesian phases in the
fenites supports evidence for a high XH2O of the fluid.
Fig. 16.2 Diagrams illustrating the main chemical changes of syenite protoliths caused by equilibration with
fenitizing fluids emanating from the carbonatite magma (Y axis: results of the mass balance calculations compiled in
Table 16.2).
In order to constrain the grade of fenitization of the anorthosites and syenites, the
chemical fenitization index (F.I.) has been determined from the quantitative mass transfer
calculation (F.I. =sum of added components). Following Kresten (1988) the fenitization grade
can be subdivided into (1) F.I. < 6.5: low-grade fenitization, (2) F.I. of 6.5-12: medium-grade
fenitization and (3) F.I. > 12: high-grade fenitization. Following this, both the fenitized wallrock
anorthosite (F.I. = 14.9) and the fenitized anorthositic xenoliths of the carbonatitic breccia (F.I. =
14.3-20.6) can be classified as high-grade fenites, whereas the F.I. of the fenitized syenites (F.I.
= 5.7-8.9) indicates low- to medium-grade fenitization, in good accordance with the lack of
metasomatically formed sodalite in the latter. Regarding the sodium to potassium ratios of the
fenitized anorthosites and syenites related to their respective protolith compositions
(Na2O/(Na2O + K2O) fenite/protolith; Fig. 16.3a) the fenitized anorthosites show an intermediate
fenitization trend, lying in between the sodic and potassic trends (subdivision of fenitization
trends by Kresten, 1988), whereas the fenitized syenites are characterized by a sodic fenitization
trend. The different fenitization trends of the anorthosites and syenites may reflect chemical
279
gradients in the metasomatic aureoles, with sodium-rich fluids migrating into the bordering
syenites (low- to medium-grade fenites), leaving back a K-enriched fluid. However, the apparent
sodic and potassic trends may also simply mirror differences in the chemical compositions of the
respective precursor rocks. Figure 16.3b displays the oxidation ratios (XFe2O3) of the fenites
plotted against the fenitization index, reflecting the oxygen fugacity conditions during their
formation. The oxidation ratios of the investigated samples are intermediate (low- to medium
grade fenitized syenites) to low (high-grade fenitized anorthosites), hence suggesting that the
fenitizing fluids were characterized by comparably low oxygen fugacities near QFM buffer
conditions.
1.0
Fenitized anorthosite
Xenoliths of fenitized
anorthosite
Fenitized syenite
1.8
1.6
sodic
1.4
0.9
0.8
0.7
1.2
1.0
XFe2O3
Na2O/(Na2 O+K2O) fenite/protolith
2.0
intermediate
0.8
0.6
0.6
0.5
0.4
0.3
potassic
0.4
a
0.2
0.0
0.2
low-grade
0.1
b
mediumgrade
high-grade and contact
0.0
0
2
4
6
8
10
12
14
16
18
20
22
24
0
2
4
6
F.I. (fenitization index)
8
10
12
14
16
18
20
22
24
F.I. (fenitization index)
Fig. 16.3: Diagrams illustrating the variation of the fenitization index (F.I.) with progressive fenitization of
anorthositic and syenitic protoliths. a) Fenitization index vs. protolith-normalized alkali ratios. b) ƒO2 variation
during progressive fenitization as reflected by the oxidation ratio of the fenites.
16.3 Origin and nature of the fenitizing solutions
Regarding the field-relationships of the investigated fenites, their metasomatic alteration appears
to be directly linked to the intrusion of the ankeritic carbonatites since (1) the fenites occur
within or in the direct neighbourhood of large carbonatite dykes and (2) no other rock type with
the potential to cause large-scale fenitization has been observed in the studied area so far. In
general terms, however, fenitizing reactions may be promoted by fluids emanating from both
alkaline and carbonatite melts, that produce distinctive patterns of fenitization, differing in XCO2,
aSiO2, aAl2O3, aCaO, XMg, ƒO2 and T (e.g. Morogan, 1994). Since the possibility, that the
fenitization at Swartbooisdrif was caused by another, not exposed magma source can not be
safely ruled out, the obtained fenitization patterns were compared to the data of ijolite- and
280
carbonatite-related fenites of Oldoinyo Lengai, Tanzania, Fen, Norway, and Alnö, Sweden
(compiled by Morogan, 1994). To summarize, fenitizing fluids related to ijolitic magmas are
characterized by low XCO2, high activities of silica, magnesium, aluminium and alkalis, low
aCaO, and high ƒO2 near the Hem-Mag buffer. Temperature gradients with distance to the
magmatic source are shallow, falling gradually from ~650°C to ~500°C. In contrast, fluids
producing carbonatitic-type fenitization have high XCO2, low activities of silica and aluminium,
and high activities of Ca. With respect to their alkali-, magnesium and iron-contents the fluids
are quite variable. Temperatures and ƒO2 are generally initially high (T: ~ 700-800°C), but
decrease sharply with distance to the magmatic source (T: <500°C), even though ƒO2 may also
increase.
Fenitization
Locality
Carbonatitic-type fenitization
Ijolitic-type fenitization
Presumably ferrocarbonatiterelated
(this study)
Oldoinyo Lengai
Hörningsholm area, Alnö
anorthositic and syenitic fenites
Baräng and Stornäset areas, Alnö
Melteig area, Fen
of Swartbooisdrif, Namibia
high: CO2 > H2O
low: H2O > CO2
relatively low: H2O > CO2
abundant calcite
scarce calcite
minor ankerite
Holla, Ulefoss and Ouarry areas, Fen
XCO2
rare hydrous mafic phases
abundant hydrous mafic phases
abundant biotite
Or and perthite in high-grade fenite
Ab and perthite in high-grade fenite
Ab and Sdl in high-grade fenite
low
high
low
no quartz present
quartz even present in HGF
no quartz present, abundant Sdl
aAl2O3
low
high
low
aCaO
very high
relatively low
relatively low
FeO/MgO
high
low
relatively high
Na2O/K2O
variable
sodic, potassic and intermediate
trends
Na2O > K2O
Na2O > K2O
intermediate to sodic trend
intermediate to sodic trend
variable
high
intermediate
aSiO2
fO2
Temperature
~700-800°C to < 500°C
~650°C to ~500°C
~770°C to ~530°C
steep T gradients
shallow T gradients
narrow aureoles; steep T gradients
Table 16.3: Main factors controlling the type of fenitization (modified after Morogan, 1994; data for Oldoinyo
Lengai after Morogan (1982) and Morogan & Martin (1985), data for Fen after Kresten & Morogan (1986) and
Kresten (1988), data for Alnö after Morogan & Woolley (1988) and Morogan (1989)).
281
Remarkably, the Swartbooisdrif fenites do not fit exactly in either of the schemes (Table
16.3). Like the carbonatitic-type fenites, they are severely desilicated and display relatively low
aAl2O3 and relatively high FeO/MgO ratios (as evident from the biotite composition, see Chapter
14.10) and reveal high higher temperatures of fenitization (see Chapter 19). It can thus be
concluded, that the fluids were definitely not derived from a silicate magma like ijolite or
nephelinite. However, regarding the low aCaO (absence of Ca-bearing hydrous mafic silicates,
losses of Ca during fenitization) and low XCO2 (abundance of hydrous mafic silicates, presence
of albite even in the high-grade fenites) of the fenites, they also differ from most of the yet
described carbonatite-related fenites. It has to be mentioned, however, that the latter two
discrimination factors, i.e. (1) XCO2 and (2) aCaO, must not imply that the fenitizing fluids are
definitely not related to the ferrocarbonatite magmatism: (1) As has been mentioned by several
authors (e.g. Wyllie & Tuttle, 1960; Currie & Ferguson, 1971; Andersen, 1986; Gittins, 1989),
carbonatite fluids may have quite variable CO2:H2O ratios, thus implying that CO2 is neither the
sole nor necessarily the dominant component of the fluids. Water saturation of carbonates may
occur at various stages of the carbonatite evolution, with the solubility of water in the carbonatite
magma increasing with increasing alkali-contents and thus with increasing fractionation (e.g.
Gittins, 1989). (2) Judging from the abundance of sodalite in the fenitized anorthosites of the
Swartbooisdrif area, the Na-rich aqueous fluids responsible for the metasomatism additionally
carried substantial amounts of Cl, thus suggesting that they represent supercritical brines. Such a
brine should generally be capable of removing alkalis, Fe, Mg, and Ca from the carbonatite
magma as chlorides, thus leading to an enrichment of these elements in the fenites. The apparent
depletion in Fe, Mg and Ca of the fenitizing fluids investigated in this study, when compared to
the bulk composition of the carbonatite matrix of the carbonatitic breccia, implies that these
elements were accumulated in readily fractionated minerals of the carbonatites such as biotite
and ankerite, hence suggesting that fractionation was already active when the fluids were
expelled.
To summarize, the qualitative retention series of the fenitizing fluids, as estimated from
the calculated compositional changes of the Swartbooisdrif fenites (i.e. H2O > Na = Cl > (Zn, Sr,
Ba, Nb, Ce, Nd, V,CO2) > (Ti, Fe, Mg, Ca, K, Al)), implies that the fenitization was caused by
late-stage fluids emanated from highly fractionated carbonatites. Consequently, the
Swartbooisdrif ferrocarbonatites represent a potential source of fenitization.
282
17 Fluid inclusion studies
17 Fluid inclusion Studies
17.1 Microthermometric studies
During the past thirty years it has become widely recognized, that the extensive fenitization of
wall-rocks, surrounding magmatic carbonatite centres, is caused by the injection of large
amounts of aqueous fluids released during the emplacement of the carbonatites (e.g. Currie &
Ferguson, 1971; Rankin, 1975; Andersen, 1986; Kresten & Morogan, 1986; Wyllie, 1989;
Gittins et al., 1990; Morogan, 1994; Morogan & Lindblom, 1995; Samson et al., 1995a, 1995b;
Bühn & Rankin, 1999; Williams-Jones & Palmer, 2002). In order to develop a model for the
hydrothermal systems around carbonatites it is important to gain knowledge about the chemical
composition of the fluid responsible. As was suggested by Dawson (1989) and Wyllie (1989)
these fluids may be capable of removing significant amounts of alkali-metals from the
carbonatite magmas, thus accounting for the close association of alkali-poor Ca-Mg carbonatites
with sodic or potassic fenites. Carbonatite fluids of quite variable compositions have been
analysed, i.e. (1) simple aqueous inclusions (2) complex aqueous inclusions containing a variety
of daughter minerals and (3) carbonic or mixed H2O-CO2 inclusions (e.g. Roedder, 1973;
Rankin, 1975; Nesbitt & Kelly; 1977; Andersen, 1986; Bühn & Rankin, 1999). In general, these
fluids are thought to evolve from a homogeneous, low-salinity carbonic into a heterogeneous
fluid, i.e. (1) a high-salinity brine and (2) a low-salinity carbonic fluid. (Andersen, 1986;
Morogan & Lindblom, 1995; Williams-Jones & Palmer, 2002). However, none of the above
mentioned studies gives qualitative or quantitative data on the composition and evolution of
fluids emanated by late-stage ferrocarbonatite melts.
This study presents the first data on the nature of orthomagmatic fluids that were derived
from ferrocarbonatites. In order to characterize the chemical composition and evolution of these
mineralising fluids, which caused the formation of the Swartbooisdrif sodalite occurrences, three
samples of the carbonatitic breccia (REE-rich carbonatitic breccia: Ku-98-130a, Ku-98-130b;
Sdl-rich carbonatitic breccia: Ku-98-131) and one quartz-syenite sample (Ku-98-40) were
chosen for microthermometric and SRXRF studies. A description of the analytical techniques is
given in the Appendix A.2.8 and A.2.9, respectively. Within the carbonatite and syenite samples
four genetically different types of fluid inclusions can be distinguished, i.e. presumably primary
inclusions of aqueous brines in sodalite and ankerite from the carbonatitic breccia, secondary
inclusions of aqueous brines in quartz of a quartz-syenite and texturally later, CO2-rich
283
inclusions in quartz of the quartz-syenite. In addition, aqueous fluid inclusions have been
observed in apatite of the carbonatitic breccia, but were too small (<0.5-2 µm) to be analysed
with the available analytical facilities. Microthermometric fluid inclusion data were processed
using FLINCOR (Brown, 1989). Results of the microthermometric investigations are listed in
Tables A.5.6.1 to A.5.6.4 in the Appendix A.5.6.
17.1.1 H2O-rich inclusions
17.1.1.1 Fluid inclusions in sodalite
The H2O-rich inclusions in sodalite of the carbonatitic breccia (Ku-98-130a, Ku-98-130b, Ku98-131) are either arranged in trails, groups or clusters or occur as solitary, randomly distributed
inclusions, without a preferred orientation of their long axes. The fact, that inclusion-trails never
cross the sodalite grain-boundaries is taken as evidence for their primary origin (Fig. 17.1a).
Moreover, when occurring close to a fracture, the inclusions are empty. The fluid inclusions
exhibit a distinct variability in both size and shape. Large inclusions (15-21 µm) predominantly
show amoeboidal shapes (Fig. 17.1 a) whereas the more common smaller ones (3-10 µm) exhibit
an either rounded or irregular shape (Fig. 17.1 b). The fluid inclusions consist of at least two
phases, an aqueous liquid and a gas-bubble with a volume ratio of 0.9-1.0; small cubic istotropic
daughter crystals (<1 µm; presumably halite) have been observed in just a few inclusions.
Fig. 17.1: Microphotographs of H2O-rich inclusions hosted by sodalite from the carbonatitic breccia. a) Large
inclusions display an amoeboidal shape, whereas b) the smaller ones are predominantly oval shaped.
The initial melting in the H2O-rich inclusions (eutectic temperatures; Te) occurred at
temperatures of -32.7° to -23.1°C, close to the metastable eutectic of the NaCl-KCl-H2O system
284
(Roedder, 1984). The salinity of the aqueous inclusions was calculated from the final melting
temperatures ( Tmf ) which range from –22.5°C to –15.0°C ( Fig. 17.2b ). Conformably, the
N 6
a
Sodalite (N=24)
N
18
Sodalite (N=47)
b
16
5
14
4
12
10
3
8
2
6
4
1
2
0
0
N
125 150 175 200 225 250 275 300 325 350 375 400 425
Th (°C)
5
Ankerite (N=7)
c
-25
-20
-15
-10
-5
0
Tmf (°C)
N
d
Ankerite (N=7)
4
4
3
3
2
2
1
1
0
0
125 150 175 200 225 250 275 300 325 350 375 400 425
N
9
8
Th (°C)
Quartz
(N=41)
e
-25
N 11
10
-20
-15
-10
Tmf (°C)
-5
Quartz (N=42)
0
f
9
7
8
6
7
5
6
4
5
4
3
3
2
2
1
1
0
0
125 150 175 200 225 250 275 300 325 350 375 400 425
Th (°C)
-25
-20
-15
-10
Tmf (°C)
-5
0
Fig. 17.2: Variation of the homogenization temperatures (Th) and the temperatures of final melting (Tmf) of
aqueous fluid inclusions. a) and b) Aqueous fluid inclusions hosted by sodalite from the carbonatitic breccia. c) and
d) Aqueous fluid inclusions hosted by ankerite from the carbonatitic breccia. e) and f) Aqueous fluid inclusions
hosted by quartz from the quartz-syenite.
resulting salt concentrations calculated after Bodnar (1993) are quite variable but high, ranging
between 19-30 wt.% NaCl equivalent (eq.) (average: 22.7 wt.%), thus indicating that the
fenitizing fluid, released by the carbonatite melt was an alkaline brine. No phase transformation,
285
indicating the presence of CO2 was observed in the temperature interval of –80°C to the critical
point of CO2. Densities of 0.86-0.98 g/cm3 have been calculated after Zhang & Franz (1987).
Homogenization of fluid inclusions in sodalite takes place at temperatures varying in a range of
270-370°C with a well defined maximum at 320-330°C (Fig. 17.2a). Hereby, fluid inclusions
hosted by sodalite of a REE-poor sample of the carbonatitic breccia (Ku-98-131) display a more
restricted range of Th (320-350°C), when compared to those in sodalite of the REE-rich
carbonatitic breccia (Ku-98-130a, Ku-98-130b; 270-320°C and 270-370°C, respectively).
A principal problem for the microthermometric analyses of the investigated fluid
inclusions in sodalite is, that cracks in sodalite tend to propagate fast especially during heating,
mostly leading to a sudden fragmentation of the thick sections into small particles. Thus,
homogenization temperatures were determined for only 24 of 48 fluid inclusions, which
remained unaffected during heating.
17.1.1.2 Fluid inclusions in carbonate
H2O-rich fluid-inclusions in carbonates from the carbonatitic breccia are generally rare, due to
extensive and repeated recrystallisation of carbonate. If present (Ku-98-130b), the fluid
inclusions are mainly hosted by granular and mostly recrystallised ankerite (Fig. 17.3 a) but are
also found in calcite. They show a tendency to form trails, which, however, never cross the
carbonate grain boundaries and thus appear to be pseudosecondary (Fig. 17.3 b). The fluidinclusions in ankerite of up to 18 µm (3-18 µm) in length are aligned and rather irregular to oval
in shape (Fig. 17.3a). At room temperature, they contain a liquid phase (H2O; 95 vol.%), a
vapour phase (H2O; 5-6 vol.%) and, rarely, an isotropic cubic phase (presumably halite).
First melting of ice (Te) in fluid inclusions hosted by ankerite occurs in the interval
between –7.0°C to -8.0°C (except one value of –29.3°C of an isolated fluid inclusion hosted by
non-recrystallised ankerite). The temperatures of final melting of ice (Tmf) commonly vary from
–5.6°C to -2.7°C (-20.7°C; Fig. 17.2d) and correspond to comparably low salinities of 4.5-6.3
wt.% NaCl eq. (22.8 wt.% NaCl eq.), thus indicating, that alkali-loss of the fluid, caused by
previous sodalite formation, predated carbonate recrystallisation. Densities range between 0.88
g/cm3 and 0.94 g/cm3 (low-salinity subgroup) and 1.04 g/cm3 (high-salinity inclusion).The liquid
and vapour phases of the fluid homogenize to liquid at significantly lower temperatures (170220°C and 200°C, respectively) than fluid inclusions in sodalite (Fig. 17.2c).
286
Fig. 17.3: Microphotographs of H2O-rich inclusions hosted by carbonates from the carbonatitic breccia. a) Trail of
pseudosecondary aqueous fluid inclusion in ankerite. b) Inclusion-trail in calcite.
17.1.1.3 Fluid inclusions in quartz
H2O-rich 2-phase (liquid + vapour) inclusions of 4-30 µm in diameter are found in granular,
interstitial quartz of a quartz-syenite sample. The inclusions are commonly arranged in trails
(Fig. 17.4 a), frequently crossing the quartz grain boundaries and are thus suggested to be
secondary in origin. They are oval to irregularly in shape with a length of up to 29 µm (2-29 µm;
Fig. 17.4 a). Small daughter crystals of halite may occur in single inclusions.
Based on their physical properties the aqueous fluid inclusions in quartz were subdivided
into three subgroups, each forming individual inclusion trails. Temperatures obtained for the
initial melting of ice (Te) fall in the range of: (1) Te: -33.9°C to –24.0°C, (2) Te: -31.4 to –22.1
and (3) Te: –25.0°C to –11.6°C, suggesting a high compositional variability of the fenitizing
fluids with respect to Na, K and Ca. The Tmf values vary between (1) –22.7°C and –17.1°C, (2)
–22.7°C to –16.6°C and (3)–13.1°C and –5.8°C (Fig. 17.2f), corresponding to salinities of 20-24
wt.% NaCl eq., 20-24 wt.% NaCl eq. and 9-17 wt.% NaCl eq., respectively. Densities of (1)
1.06-1.11 g/cm3, (2) 0.99-1.06 g/cm3 and (3) 0.94-1.00 g/cm3 have been calculated for the three
different populations of fluid inclusions. Liquid and vapour homogenize at temperatures of (1)
140-170°C, (2) 190-240°C and (3) 180-210°C, which are generally lower than the Th observed
for aqueous inclusions in sodalite and in the range of the Th of aqueous fluid-inclusions in
287
carbonates of the carbonatitic breccia (Fig.17.2e). The high variability of the Th calculated for
the fluid inclusions must be related to temperature changes during trapping of the fluid
inclusions, since the homogenization temperatures remained constant during repeated heating,
thus ruling out leakage of the inclusions.
Fig. 17.4: a) Microphotograph of secondary aqueous inclusions hosted by quartz from the quartz-syenite. b)
Microphotograph of secondary CO2-rich inclusions hosted by quartz from the quartz-syenite.
17.1.2 CO2-rich inclusions
The fourth type of fluid-inclusions, the CO2 inclusions, are restricted to quartz of the syenite,
slightly dark in appearance and can be distinguished from the H2O-rich inclusions by shape and
size. They are rather oval with a length of 4-21 µm and show no preferred orientation of their
long axis within the inclusion trails (Fig. 17.4 b). The inclusion trails generally cut both the H2Oinclusion trails and the quartz grain boundaries and are thus secondary in origin using the criteria
of Roedder (1984). CO2-rich fluid inclusions produced during the early stages of fenitization
have never been observed.
Melting temperatures of 27 microthermometrically analysed inclusions cover a restricted
range of –56.7°C to –56.5°C, with a well defined maximum at –56.6°C, the melting temperature
of pure CO2 and thus indicating that phases such as CH4, N2 and SO2, which would lower the
melting temperature, are absent or negligible (Fig. 17.5 a). Homogenization takes place into the
fluid phase at 8.3-23.7°C with two distinct maximums, one at 13-17°C and a second one at 20-
288
24°C (Fig. 17.5 b). Densities of 0.73-0.87 g/cm3 for pure CO2 have been calculated after Kerrick
& Jackobs (1981).
N 22
20
N 6
N = 27
N = 27
5
18
16
4
14
12
3
10
8
2
6
4
2
1
a
b
0
0
-57.0 -56.9 -56.8 -56.7 -56.6 -56.5 -56.4 -56.3 -56.2 -56.1 -56.0
Tmf (°C)
0
2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
Th (°C)
Fig. 17.5: Results of microthermometrical measurements of CO2-rich inclusions in quartz from the quartz-syenite.
a) Histogram showing the temperatures of final melting of CO2-rich inclusions. b) Corresponding homogenization
temperatures.
17.1.3 Discussion and interpretation
Except rare and small aqueous inclusions hosted by apatite, the primary aqueous, highly saline
fluid inclusions in sodalite (19-30 wt.% NaCl, Th: 270-370°C) constitute the earliest
recognisable generation of fluid in the carbonatitic breccia. Since textural relationships and
microthermometrical measurements display no evidence for major post-genetic changes, the
composition of the these fluid-inclusions is presumably close to the initial composition of the
fenitizing fluid. In contrast, primary to pseudosecondary aqueous fluid inclusions hosted by
ankerite of the cementing carbonatite matrix of the carbonatitic breccia display lower minimum
temperatures of entrapment (Th: 170-220°C). Rare fluid inclusions preserved in nonrecrystallised ankerite display a similar salinity like those in sodalite (23 wt.% NaCl eq.),
whereas those hosted by recrystallised ankerite reveal significantly lower salinities (4-6 wt.%
NaCl eq.), thus suggesting that the recrystallisation of ankerite post-dated the major sodaliteforming event.
Secondary fluid inclusions in quartz of the quartz-syenite are characterized by a high
variability of both the fluid composition and the Th (L-V). Since the three distinct types of fluid
populations, occurring in the three individual inclusion trails, lack clear textural relationships, it
289
was impossible to constrain their relative temporal succession. High-salinity groups (20-24 wt.%
NaCl eq.) were trapped at temperatures of 140-170°C and 190-240°C, whereas the low-salinity
group (9-17 wt.% NaCl eq.) displays Th in a higher range (180-210°C). Two possible
explanations may account for the conspicuous physical properties of the fluid inclusions, i.e. (1)
injection of a high-salinity aqueous fluid leading to an extensive albite formation under still high
temperatures (strong decreases of salinity, weak decreases of Th), followed by the crystallization
of muscovite during cooling (increases of salinity and decreases of Th) or (2) injection of a highsalinity aqueous fluid leading to albite formation, followed by the repeated injection of a highsalinity aqueous fluid at lower temperatures. By now, possibility (2) is the preferred model, since
the fluid inclusions seem to follow a continuous trend towards lower salinities, when plotted in
the Th-salinity diagram. As textural evidence for a cogenetic formation of the CO2 and the H2O
inclusions in quartz of the quartz-syenite is lacking, the late CO2 inclusion trails are ascribed to a
weak low temperature hydrothermal overprint.
1.2
30
+∆P,-∆T
+ ∆S
1.1
20
-∆S
-∆P,+∆T
3
Density (g/cm )
Salinity (wt.% NaCl eq.)
25
Sdl (Ku-98-130a)
Sdl (Ku-98-130b)
Sdl (Ku-98-131)
15
Qtz 1 (Ku-98-40)
Qtz 2 (Ku-98-40)
Qtz 3 (Ku-98-40)
10
5
1.0
0.9
Ank 1 (Ku-98-130b)
Ank 2 (Ku-98-130b)
a
b
0.8
0
100
150
200
250
Th (°C)
300
350
400
0
5
10
15
20
25
30
Salinity (wt.% NaCl eq.)
Fig. 17.6: a) Homogenization temperature (Th) vs. salinity for all inclusions as a function of inclusion type and host
mineral. b) Salinity vs. density for all inclusions (the inset after Samson et al. (1995a) qualitatively illustrates the
effects on density caused by increases or decreases salinity (S), temperature (T) and pressure (P)).
The temperature and pressure conditions at which the fluid was trapped (Fig. 17.6 a) as
well as the salinity of the fluid inclusions (Fig. 17.6 b) strongly control the densities of the
aqueous fluid inclusions. In a salinity-controlled system a strong positive correlation between
salinity and density should be expected (e.g. Shepherd et al., 1985; Samson et al., 1995a), which
indeed is recorded by the fluid inclusions hosted by ankerite from the carbonatitic breccia and by
quartz of the quartz syenites (Fig. 17.6 b). Following this, the fenitizing fluid underwent major
compositional changes which may be attributed mainly to a variable sodium-loss during the
290
formation of large amounts of Na- and NaCl-rich silicates (e.g. albite and sodalite) during the
fenitization of the anorthosites and syenites. In contrast, the almost vertical trend displayed by
the sodalite-hosted fluid inclusions in the density-salinity plot (Fig. 17.6 b) may be better
explained by temperature and/or pressure variations during sodalite formation, whereas salinitychanges of the fluid seem to be of only minor importance (Fig. 17.6 a). However, the fluid
inclusions in sodalite are generally less dense than those in ankerite and quartz, constraining that
they were formed under higher temperatures and/or lower pressures (Fig. 17.6 b).
Given the postulated temporal succession of the fluid inclusions (e.g. sodalite → quartz,
ankerite), the data indicates that the fenitizing fluids (1) led to the formation of albite and
sodalite in the fenitized anorthositic xenoliths over a certain temperature interval under almost
constant and high salinities, (2) caused the transformation of K-feldspar into albite in the
fenitized syenites under lower temperatures, leading to a continuous decrease of the salinities of
the fenitizing fluids followed by a repeated influx of high-salinity fluids and (3) were enclosed
by the recrystallising carbonates from the carbonatitic breccia, with the comparably low salinities
of the fluids reflecting previous albite and sodalite formation. The texturally late CO2-inclusions
are presumably formed during a later, independent event of low-temperature alteration. Absolute
temperature and pressure conditions of trapping will be discussed in some detail in Chapter 19.3.
17.2 High-resolution synchrotron micro-XRF analysis
Due to the small sizes (5-20 µm) of the fluid inclusions trapped in sodalite, quartz, ankerite,
calcite and apatite, the low concentrations of the elements of mineralogical interest (e.g. Sr, Nb,
Ba and REE) and the sensitiveness of sodalite regarding temperature changes, a high spatial
resolution combined with low detection limits of ppm to sub-ppm is needed for the
measurements of their respective chemical compositions. The analyses of the fluid inclusions
were performed with the standard polychromatic synchrotron radiation X-ray fluorescence set-up
at beamline L at HASYLAB, Hamburg, which allows both, the multi-element analysis of even
the high energetic REE and a destruction-free in situ measurement of the trace element
geochemistry of the fluid inclusions under almost constant temperatures. In order to calculate the
trace element compositions, solitary inclusions were selected and analysed with a beam size of
10 µm in diameter, covering the entire inclusion. The salinity of individual fluids was
determined by microthermometric fluid inclusion measurements whereas their trace element
contents were estimated from synchrotron micro-XRF analyses, which, in general, gives
291
reasonable results for elements with an atomic number > 15. EMP major element analyses (Ba,
Sr or Fe) were used for the matrix correction and the SRXRF trace element quantification and
standardisation. For the estimation of the composition of the fluid inclusions the element
contents analysed for the inclusion-free host mineral were subtracted from the respective values
of the total analyses (fluid inclusion + hosting mineral). The values calculated for the trace
elements of the fluid inclusions must thus be regarded as minimum trace-element contents. The
results for the investigation of 7 fluid inclusions in sodalite (Ku-98-14 and Ku-98-130b), 10 fluid
inclusions in quartz (Ku-98-40) and 3 fluid inclusions in ankerite (Ku-98-14 and Ku-98-130b)
are compiled in Table 17.1 (Drüppel & Rickers, in prep.).
Mineral
Sodalite
Quartz
Min
(N=7)
Ankerite
Min
(N=10)
Max
(N=7)
Max
(N=10)
Min
(N=3)
Max
(N=3)
ppm
S
K
Ca
Fe
Mn
Ti
Zn
Ga
As
Cu
Br
Sr
Ba
Nb
Sb
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Er
Yb
Lu
110
20
60
2
0
5
222
0
27
5
9
28
10
1
2
119
-
180
240
290
243
2
107
303
9
924
875
298
29
13
344
631
213
-
-
264
22
11
0
84
35
5
-
404
91
194
4
668
943
1295
-
150
6
1069
2611
1270
115
61
75
8
33
15
6
3
700
38
1688
3541
1492
218
94
166
18
121
38
29
17
Table 17.1: Trace element composition of aqueous fluid inclusions hosted by sodalite and ankerite from the
carbonatitic breccia and by quartz of the quartz-syenite as revealed by SRXRF analyses.
SRXRF analyses of apatite crystals from the carbonatitic breccia (Ku-98-14) display
similar spectra for the inclusion-free apatite matrix and the fluid inclusion-rich areas. This is
292
most probably an effect of the overall high Sr, Ba and REE contents of apatite and the
comparably small volume of the fluid inclusions with respect to the apatite host, thus making a
quantitative determination of the trace element compositions impossible.
Primary aqueous saline inclusions (19-30 wt.% NaCl eq.) trapped during the growth of
sodalite from the carbonatitic breccia (Ku-98-14, Ku-98-130b) contain appreciable amounts of
Sr (5-875 ppm), Ba (9-298 ppm), La (1-344 ppm) and Ce (2-631 ppm), in good accordance with
the results of the mass change calculations for the fenitization of the anorthosites (see Chapter
16). Minor amounts of S (110-180 ppm), Nb (28-29 ppm), K (20-240 ppm), Fe (60-290 ppm),
Mn (2-243 ppm), Zn (0-2 ppm), Ga (5-107 ppm), Sb (10-13 ppm), and Pr (119-213 ppm) may be
present as well, indicating that the fluid was also capable of transporting these elements. The
composition of the fluid inclusions is presumably close to the initial composition of the
fenitizing fluid. The fact that Ca is below the detection limit, constrains the suggestion, that Ca
was incorporated in previously crystallized carbonates, hence implying that the enclosed fluids
represent a late-stage fluid-phase of a highly fractionated carbonatite magma. Judging from the
low K contents of < 240 ppm, the fenitizing fluid is suggested to be characterized by a high Na:K
ratio.
In clear contrast, secondary aqueous inclusions (9-24 wt.% NaCl eq.) trapped in quartz
crystals of a bordering syenite (Ku-98-40) contain no La or Ce but display higher amounts of Sr
(up to 943 ppm) and Ba (up to 1295 ppm), as well as minor amounts of K (264-404 ppm), Ca
(22-91 ppm), Fe (11-194 ppm), Mn (0-4 ppm), and Ti (84-668 ppm). The composition of these
fluids points to NaCl-loss of the fenitizing fluid during previous sodalite formation, accompanied
by a relative enrichment of Sr, Ba Ca and H2O.
The H2O-rich fluid inclusions (4.5-6.3 wt.% NaCl eq.) hosted by recrystallised ankerite
from the REE-rich carbonatitic breccia (Ku-98-130b) strongly differ from the other aqueous
fluid inclusion-types in containing no Ba, Fe and K but revealing comparably high amounts of Sr
(150- 700 ppm) and the REE (ΣREE: 5265-7421 ppm). As for the carbonatite, the (La/Nd)cn
ratios are high, ranging from 1.4-2.4. The REE patterns of the fluid inclusions are straight,
except a weak negative (Sm/Sm*)cn anomaly of 0.4-0.7 (Fig. 17.7), which is also displayed by
both the REE-poor ((Sm/Sm*)cn: 0.6-0.8) and the REE-rich carbonatitic breccia ((Sm/Sm*)cn:
0.6-0.7). The fluid inclusions reveal either an accentuated negative Ho anomaly or Ho contents
below the detection limit, thus suggesting that Ho was either incorporated by a previously
crystallized mineral or remained in the carbonatite melt. When compared to the carbonatitic
293
breccia, the fluid inclusions are characterized by flatter patterns overall, suggesting previous or
contemporaneous crystallization of Sr-La-Ce carbonates. Regarding their high Sr and LREE
contents, the fluids are a potential source for the Sr and REE enrichment of the REE-rich
carbonatitic breccia. Following this, the REE-enrichment of the REE-rich carbonatitic breccia
may better be explained by the influx of late-stage, highly evolved carbonatite fluids than by
extensive fractionation of the carbonatite melt.
1000000
Aqueous fluid inclusions in ankerite
sample / chondrite
100000
10000
1000
100
10
1
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
Fig. 17.7: Chondrite-normalized rare-earth element patterns of individual aqueous fluid inclusions in ankerite from
the carbonatitic breccia. Chondrite-normalized REE-patterns of the REE-poor carbonatitic breccia (dark grey) and
the REE-rich carbonatitic breccia (light grey; whole-rock ICP-AES analyses) are illustrated for comparison.
17.3 Implications for the evolution of the fenitizing fluids
This study was initiated in order to provide physicochemical data on the orthomagmatic fluids
emanated by the Swartbooisdrif carbonatites. The data indicate that the fluids, even those
enclosed in early crystallized sodalite from the carbonatitic breccia, are aqueous fluids
containing no CO2. Orthomagmatic CO2-inclusions or low-salinity carbonic fluids, that are
generally expected to be produced during the early stages of carbonatite evolution, have never
been observed, thus constraining that carbonate fractionation was already active when the fluids
were expelled. Following this, the aqueous fluid inclusions can be described as late-stage
carbonatite fluids.
294
Aqueous fluid inclusions in sodalite from the carbonatitic breccia contain high
concentrations of Na, Sr, Ba, Fe, Nb, and the LREE and may thus be responsible for the
observed compositional changes of the anorthosites during fenitization (see Chapter 16). The fact
that K is present in only minor proportions strongly suggests that the fluid is NaCl-dominated, in
good accordance with the sodic fenitization type (i.e. the formation of large amounts of albite
and sodalite). Since implications for post-genetic changes are lacking (e.g. necking down or
decrepitation phenomena), the composition of the fluids is presumably close to the initial
composition of the carbonatite fluid, even though previous and contemporaneous sodalite
crystallization may have caused substantial NaCl-loss of the fluid before trapping.
Presumably later, H2O-rich fluid inclusions hosted by recrystallised ankerite from the
carbonatitic breccia strongly differ from those in sodalite in containing distinctly lower amounts
of Na and no Ba, Fe and K, thus suggesting that these element were incorporated in previously
crystallized minerals like albite, sodalite, ankerite, magnetite and biotite, which were formed
during the progressive fenitization of the wall-rocks and the crystallization of the carbonatite
itself. Decreases of Na, K and Fe in the fluid may in turn have caused a relative increase in the
remaining fluid species, such as Sr and the REE of the ankerite-hosted fluid inclusions. The fact,
that these elements are highly concentrated in the REE-rich carbonatitic breccia, provides
evidence that the REE-enrichment was caused by autometasomatic processes.
The secondary aqueous inclusions trapped in quartz crystals of quartz-syenite seem to
preserve an intermediate stage of the fluid evolution. The highly variable composition of these
fluids points to previous and contemporaneous Na-loss of the carbonatite-released fluid during
albite and sodalite formation, accompanied by a relative enrichment of K, Ca, Fe, Ti, Sr and Ba.
It has to be mentioned, however, that the occurrence of a low-temperature, high-salinity group of
fluid inclusions testifies to repeated injections of high-salinity carbonatite fluids, thus making it
impossible to unravel the exact chemical evolution of the fluids preserved in the quartz-syenite.
As a conclusion, the data imply, that the fluids preserved in sodalite and ankerite from the
carbonatitic breccia and in quartz of the quartz-syenite represent direct samples of highly
evolved, alkali-carbonatitic fluids, expelled from a fractionated carbonatite magma. Major
compositional changes of the fluid are mainly caused by interrelated chemical reactions between
the fluid and the lithologies bordering the carbonatite centres.
18 Stable isotope composition
295
18 Stable Isotope Composition
18.1 Oxygen and Carbon Isotopic Composition
Stable isotopes can serve as a powerful tool for constraining the source rocks of carbonatites and
their magmatic and post-magmatic evolution (e.g. Deines, 1989). In general, variations in the O
and C isotopic ratios in carbonatite complexes are mainly attributed to (1) the isotopic
composition of the magma sources, (2) fractionation processes during the carbonatite evolution
(3) isotopic composition of the country rock and (4) post-magmatic alteration (see Deines, 1989,
for a review). The range of the δ18O and the δ13C compositions of primary igneous carbonatites
was first defined by Taylor et al. (1967) and called "carbonatite box", ranging between 6 and
10‰ for δ18O and between -4 and –8‰ for δ13C. Recently, however, Keller & Hoefs (1995)
supposed a more restricted composition for the "primary" isotope signatures of carbonatites
("primary" in this case means unaffected by secondary alteration processes), based on analyses
of the stable isotope patterns of completely fresh natrocarbonatite from Oldoinyo Lengai (-6.3 to
–7.1 ‰ δ13C and 5.5-7.0 ‰ δ18O), which is adopted for this study. Remarkably, only part of the
yet investigated carbonatites fall in the above mentioned compositional fields, whereas a
considerable number of samples, especially the ankerite carbonatites, have significantly higher C
and O ratios, which are generally thought to be caused by secondary processes (e.g. Deines,
1989; Horstmann & Verwoerd, 1997).
This study addresses the O and C isotope signatures of the Swartbooisdrif carbonatites as
well as their influence on the O isotope patterns of the bordering syenites and anorthosites in the
light of magmatic-metasomatic as well as hydrothermal or meteoric water alteration (Drüppel et
al., subm.). The δ18O and δ13C data of separated minerals from carbonatites (ankerite, magnetite,
biotite), anorthosites (plagioclase), fenitized syenites (albite) and the fenitized nepheline syenite
(K-feldspar, nepheline, biotite) are listed in Table A.5.5.1 in the Appendix A.5.5. A description
of the analytical techniques is given in the Appendix A.2.6.1.
18.1.1 Oxygen isotopic composition of anorthositic rocks
The oxygen isotopic composition of the anorthositic rocks of the Kunene Intrusive Complex has
been discussed in some detail in Chapter 8.1. In order to investigate the influence of carbonatite-
296
related fluids on the anorthositic country rocks plagioclase was separated from 4 dark
anorthosites sampled in the vicinity of the sodalite-mining area and from 4 dark anorthosites
sampled in distinct parts of the anorthosite massif. The fact, that anorthosite samples from both
areas display similar and magmatic δ18O ranges for plagioclase (5.94 ± 0.21 ‰ and 5.82 ± 0.15
‰, respectively) is taken as evidence, that the circulation of carbonatite-derived metasomatic
fluids was confined to contacts between the carbonatites and their wallrocks and reached no
large areal extent. The lack of broad metasomatic aureoles around the carbonatite dykes implies
that the Swartbooisdrif ferrocarbonatites are deep-seated intrusions, emplaced in crustal levels
where the effects of metasomatism are less prominent.
18.1.2 Oxygen isotopic composition of fenitized syenites and nepheline syenites
The oxygen isotopic ratios of orthomagmatic K-feldspar and plagioclase of the non-altered
syenites of the Swartbooisdrif area are discussed in Chapter 8.2. In addition, metasomatically
formed albite was separated from two samples of the fenitized syenite (Ku-98-70s, Ku-99-19s).
Remarkably, the δ18O values of feldspar separates of the non-altered syenite and the fenitized
syenite are quite similar, ranging between 7.20-7.92 ‰ (average: 7.45 ± 0.24 ‰) and 7.53-7.77
‰ (average: 7.65 ± 0.12 ‰), respectively (Fig. 18.1). Nepheline of the fenitized nepheline
syenite (Ku-99-15a) exhibits a slightly lower δ18O of 6.93, whereas the K-feldspar from the same
6
6
K-feldspar, syenite
plagioclase, syenite
5
4
4
3
3
2
2
1
1
a
0
-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
18
δ OSMOW
K-feldspar, Ne-syenite
abite, fenitized syenite
nepheline, Ne-syenite
5
0
b
-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
18
δ OSMOW
Fig. 18.1: Oxygen isotopic composition of hand-picked feldspar and nepheline separates from the syenites, fenitized
syenites and the fenitized nepheline syenite. a) Oxygen isotope data of orthomagmatic feldspar from the syenites. b)
Oxygen isotope data of metasomatically formed albite from the fenitized syenites and of K-feldspar and nepheline
from the nepheline syenite.
297
sample has a δ18O of 7.62 ‰, similar to those of syenite feldspars (Fig. 18.1b). Orthomagmatic
biotite of the nepheline syenite sample shows a δ18O of 1.73 ‰.
The homogeneity of the δ18O values of feldspars and nepheline from the different rocktypes is striking, since the two samples of the fenitized syenite and the nepheline syenite sample
display clear textural and chemical evidence for the interaction with fenitizing fluids. Two
explanations may account for the obtained results: (1) the carbonatitic fluids had no influence on
the δ18O of the feldspars and nepheline, or (2) the metasomatic fluid had magmatic temperatures
and 18O/16O ratios similar to or only slightly higher than those of the syenites. By now, model (2)
is favoured, as albite in the alkali-syenites and the sodium-rich rims of nepheline in the nepheline
syenite are not magmatic in nature but were formed during the metasomatic event. Following
this, the fenitizing fluids emanated by the carbonatites should have had δ18O values of ≥ 7.5-7.8
‰ and magmatic temperatures of c. 600-890°C (see Chapter 9.2).
18.1.3 Oxygen and carbon isotopic compositions of the carbonatitic breccia and the
Separated magnetite grains from the carbonatitic breccia (Ku-97-26, Ku-99-02, Ku-99-SA11)
and from the ferrocarbonatite vein (Ku-01-03) reveal variable values of δ18O, in the range of –
4.21 to –1.46 ‰ and of –2.63 ‰, respectively. Biotites from the carbonatitic breccia (Ku-98-48,
Ku-99-5a) show 18O values of 2.77-3.41 ‰ (Fig. 18.2).
Separated ankerite grains from 4 samples of the carbonatitic breccia (Ku-97-26, Ku-9902, Ku-99-SA8, Ku-99-SA11) cover a narrow range in δ13C (-6.86 to -6.98) and exhibit a
stronger variation in δ18O (8.95 to 9.73) Similar but slightly lower values are displayed by
ankerite from a late ferrocarbonatite vein (Ku-01-03; δ13C: -6.73, δ18O: 8.91; Fig. 18.2), which
agree well with the values obtained by Thompson et al. (2002) for similar samples. The
13
C
contents of carbonates from both carbonatite generations are in good agreement with the δ13C
values of the fresh Oldoinyo Lengai natrocarbonatite lava and its carbonate phenocrysts (-6.3 to
–7.1 ‰ δ13C) given by Keller & Hoefs (1995). In contrast, the 18O values of the Swartbooisdrif
dykes fall into the carbonatite box of Taylor et al. (1967) but are distinctly higher than those
reported for unmodified partial mantle-melts from Oldoinyo Lengai (5.5-7.0 ‰ δ18O; Keller &
298
Hoefs, 1995), thus indicating the involvement of magmatic and/or hydrothermal phases during
their petrogenesis (Fig. 18.3).
6
5
Carbonatitic breccia
and ferrocarbonatite
veins
FREQUENCY
Ankerite
4
Biotite
Magnetite
3
2
1
0
-6 -5 -4 -3 -2 -1 0
1
2
3
4
5
6
7
8
9 10 11
18
δ OSMOW
Fig. 18.2: Oxygen isotopic composition of hand-picked magnetite, biotite and ankerite separates from the
carbonatitic breccia and the ferrocarbonatite veins.
Following Deines (1989) enrichment of δ18O of carbonatites may be due to: (1)
emanation of fluids during pressure reduction at the time of emplacement (e.g. preferential
escape of
18
O-depleted water or increase of the H2O:CO2 ratio in the remaining fluid), (2)
exchange with magmatic, high 18O fluids (e.g. carbonatite fluids with a low H2O:CO2 ratio), (3)
retrograde exchange with magmatic fluids at temperatures of c. 100°C, (4) retrograde exchange
with 18O-rich hydrothermal fluids, displaying high H2O:CO2 ratios and (5) the influx of meteoric
water at temperatures below 200-250°C, which previously had undergone isotopic exchange
with 18O-rich sedimentary or metasedimentary rocks at low temperatures. Both, point (1) and (5)
after Deines (1989) may generally account for the elevated
18
O values of the ferrocarbonatites
investigated in this study. However, as has been demonstrated in Chapter 8.1, isotopic exchange
between anorthosites and meteoric waters led to a decrease of the 18O values of the anorthosites,
thus suggesting that the circulating meteoric solutions in the investigated study area, at least
those altering the anorthosites, rather had
18
O-poor compositions. Point (5) seems to be a
plausible explanation for the 18O patterns of the investigated samples, since the oxygen isotopic
compositions of metasomatically formed albite and nepheline from the syenites and the
nepheline syenites point to the influx of fenitizing fluids with comparably low 18O values of 7.57.8 ‰, thus being indeed
18
O-depleted with respect to ankerite from the carbonatites (8.91 to
9.73 ‰ δ18O). However, regarding clear evidence for low-temperature alteration of the studied
299
samples (abundance of late CO2-inclusions in quartz of a quartz-syenite, see Chapter 17.1.3), a
combination of process (1) and (5) seems to be most likely.
-3
ankerite from carbonatitic breccia
(Swartbooisdrif)
ankerite from ferrocarbonatite dikes
(Swartbooisdrif)
fresh natrocarbonatite lava
(Oldoinyo Lengai; Keller &
Hoefs, 1995)
phenocrysts from fresh
natrocarbonatite (Oldoinyo Lengai,
Keller & Hoefs, 1995)
-4
13
δ CPDB
-5
“primary igneous
carbonatites”
(Taylor et al., 1967)
mantle field defined
by oceanic basalts
(Deines, 1989)
-6
-7
-8
primary igneous
carbonatites
(Keller & Hoefs, 1995)
-9
4
5
6
7
8
9
10
11
18
δ OSMOW
Fig. 18.3: δ13C-δ18O plot of hand-picked ankerite separates from both the carbonatitic breccia and the
ferrocarbonatite veins of the Swartbooisdrif alkaline province (dotted-line box: based on C/O isotopes of fresh
natrocarbonatite lava and phenocrysts (Keller & Hoefs, 1995), solid-line box: based on C/O isotopes in basalts
(Kyser, 1986; Nelson et al., 1988; Deines, 1989), dashed-line box: based on C/O isotopes of carbonatites (Taylor et
al., 1967)).
Regarding their oxygen and carbon isotopic composition, the Swartbooisdrif carbonatites
differ significantly from most other ankerite and ankerite/calcite carbonatites, which are mostly
characterized by variable and high
18
O and
13
C ratios (e.g. Deines, 1989; Horstmann &
Verwoerd, 1997), lying outside the primary igneous carbonatite field as defined by Taylor et al.
(1967; Fig. 18.4). The ankerites of these carbonatites are generally thought to be of secondary
origin, with their formation being caused by the influx of Fe- and Mn-rich meteoric waters (e.g.
Horstman & Verwoerd, 1997). Only two ankerite carbonatites, i.e. the ferrocarbonatites of
Marinkas Quellen, Namibia (Verwoerd, 1993; Horstmann & Verwoerd, 1997), and the Barra do
Itapirapuã ferro- to magnesiocarbonatites, Brazil (Andrade et al., 1999), have stable isotope
ratios adjacent to those of the Swartbooisdrif carbonatites, plotting within the magmatic C-O
field of Taylor et al. (1967; Fig. 18.4). Remarkably, however, oxygen and carbon isotope values
300
of calcio- and magnesiocarbonatites of the Lupongola complex, SW Angola (Alberti et al, 2000),
are quite similar to those obtained for the Swartbooisdrif. Following these arguments, the stable
isotope values obtained for the Swartbooisdrif ferrocarbonatites seem to be quite exceptional, as
13
δ CPDB
they are concordant with a magmatic origin.
4
3
2
1
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
Lupongola
calciocarbonatite
magnesiocarbonatite
Swartbooisdrif (this study)
ferrocarbonatite
Barro do Itapirapua
Goudini
Kruidfontain
Marinkas Quellen
Osongombo
Spitskop
4
6
8
10 12
14 16
18 20 22 24
18
δ OSMOW
Fig. 18.4: δ13C-δ18O plot of hand-picked ankerite separates from both the carbonatitic breccia and the
ferrocarbonatite veins of the Swartbooisdrif alkaline province and from carbonatites elsewhere in southern Africa
and Brazil (dashed-line box: based on C/O isotopes in basalts (Kyser, 1986; Nelson et al., 1988; Deines, 1989),
solid-line box: based on C/O isotopes of carbonatites (Taylor et al., 1967); isotopic data of ankerite and
ankerite/calcite carbonatites from Goudini, Kruidfontain, Marinkas Quellen, Osongombo and Spitskop: Horstmann
& Verwoerd (1997), isotopic data of magnesio- to ferrocarbonatites from the Barro do Itapirapuã complex, Brazil:
Andrade et al. (1999), isotopic data of calcio- and magnesiocarbonatites of the Lupongola complex, Angola: Alberti
et al. (2000).
18.2 Sulphur isotopic composition
Sulphur isotope data are yet available for relatively few carbonatites, with the analysed sulphides
including pyrite, chalcopyrite, galena and pyrrhotite. The sulphur isotope composition of mantlederived carbonatite melts can be estimated from the sulphur isotope ratios of meteorites, basic
sills and ocean-floor basalts. Following Deines (1989) most of these analyses fall in a narrow
301
range of –1 to +1 ‰, a range, which is generally thought to be characteristic of the pristine
mantle. However, significant 34S enrichments with respect to these values have been observed in
some basic intrusions, suggesting that parts of the mantle are enriched in 34S (e.g. Deines, 1989).
In contrast, the sulphur isotope composition of carbonatites is variable, ranging between –13 to
+2 ‰ for δ34S and thus deviating significantly from the isotopic ratio of the pristine mantle (see
Deines, 1989, for a review). These differences are thought to be mainly caused by (1) isotope
heterogeneity of S within the mantle or (2) changes in the O fugacity during the carbonatite
formation (e.g. Deines, 1989).
This study addresses the S isotope signatures of the carbonatitic breccia and those of the
fenitized anorthosites and syenites (Drüppel & Wagner, in prep.). In order to investigate the
changes of δ34S during the crystallization of the carbonatites, presumably early, euhedral pyrite
from the sodalite-rich carbonatitic breccia (Ku-01-OD1, Ku-01-OD2) as well as pyrite and
chalcopyrite from late-stage sulphide-oxide veins in the sodalite-rich carbonatitic breccia (Ku99-SA8) and the REE-rich carbonatitic breccia (Ku-98-130b) have been analysed for their
sulphur isotope composition. Moreover, the S isotope patterns of metasomatically formed pyrites
from a fenitized leucogabbronorite (Ku-97-03b) and a fenitized syenite (Ku-97-05a) have been
determined, to constrain the isotopic composition of the carbonatite fluids. The δ34S data of the
analysed sulphides are depicted in Table A.5.5.2 in the Appendix A.5.5 and are illustrated in Fig.
18.5. A description of the analytical method is given in the Appendix A.2.6.2.
The sulphur isotopic ratios of orthomagmatic pyrite grains, dispersed in the carbonatitic
breccia range between 3.76‰ and 5.13 ‰ (Fig 18.5a, b), being about 4 ‰ higher than the
34
S
range of the pristine mantle. These values may either reflect (1) a δ34S-enriched mantle source or
(2) fractionation governed by oxidised S species in the melt. In clear contrast, distinctly lower,
negative δ34S values have been determined for pyrite and chalcopyrite in late sulphide-oxide
veins (-0.03 to –3.25 ‰ δ34S; Fig 18.5c, d), with the lowest values obtained for pyrite and
chalcopyrite from the REE-rich carbonatitic breccia (-2.07 to –2.43 ‰ δ34S and –2.84 to –3.25
‰, respectively). Notably, late-stage chalcopyrite from the sodalite-rich carbonatitic breccia (0.03 to –1.31 ‰ δ34S) is intimately associated with magnetite/hematite, whereas the REE-rich
sample Ku-98-130b additionally contains accessory barite. Following this, the carbonatite
sulphides follow a continuous trend towards more negative δ34S values coupled with an increase
of the oxygen fugacity. This trend can be interpreted in terms of changing proportions of
oxidised and reduced S species in the melt: (1) During the early stages of crystallization the
oxygen fugacity was presumably low (formation of magnetite) whereas the proportion of the
302
reduced S species (H2S, HS-)in the melt was high, promoting the precipitation of pyrite and the
formation of almost pure, SO4-free sodalite and thus in turn leading to an increase of the oxidised
5
5
Ku-01-OD2
Pyrite
4
FREQUENCY
FREQUENCY
4
Ku-01-OD1
Pyrite
3
2
1
3
2
1
a)
0
-4
b)
0
-3
-2
-1
0
1
2
3
4
5
6
-4
-3
-2
-1
0
34
1
2
δ SCDT
4
6
Ku-98-130b
Chalcopyrite
Pyrite
4
FREQUENCY
FREQUENCY
5
5
Ku-99-SA8
Chalcopyrite
3
2
1
3
2
1
c)
0
-4
-3
-2
-1
0
1
2
3
4
5
d)
0
6
-4
-3
-2
-1
0
34
1
2
3
4
5
6
2
3
4
5
6
34
δ SCDT
δ SCDT
5
5
Ku-97-05
Pyrite
Ku-97-03b
Pyrite
4
FREQUENCY
4
FREQUENCY
4
δ SCDT
5
3
2
3
2
1
1
0
3
34
e)
-4
f)
0
-3
-2
-1
0
1
34
δ SCDT
2
3
4
5
6
-4
-3
-2
-1
0
1
34
δ SCDT
Fig. 18.5: Oxygen isotopic composition of pyrite and chalcopyrite from the carbonatitic breccia, the fenitized
anorthosites and the fenitized syenites. a) and b) Isolated pyrite grains, dispersed in the carbonatite matrix of the
carbonatitic breccia. c) and d) Pyrite and chalcopyrite occurring in texturally late sulphide veins of the carbonatitic
breccia. e) Metasomatically formed pyrite from the fenitized syenite. f) Metasomatically formed pyrite from the
fenitized anorthosite.
303
S species in the magma. (2) In contrast, both the oxygen fugacity and the proportion of the
oxidised S species (HSO4-, SO42-) continuously increased during the late stages of the carbonatite
evolution, hence leading to the precipitation of barite and consequently increasing the relative
abundances of the reduced S species in the carbonatite magma.
Remarkably, the δ34S values obtained for metasomatically formed pyrite from both the
fenitized syenites (1.17-3.40 ‰ δ34S; Fig. 18.5e) and the fenitized anorthosites (3.27-3.40 ‰
δ34S; Fig. 18.5f) are similar but slightly lower than those of early crystallized pyrite from the
carbonatitic breccia, hence implying depletion of δ34S in the carbonatite fluid with respect to the
carbonatite melt. Following this, it may be possible that (1) the release of the fenitizing fluids
predated the formation of pyrite in the carbonatitic breccia, thus preserving lower δ34S contents
of the carbonatite magma or (2) the fluid was characterized by a higher oxygen fugacity when
compared to the carbonatite melt.
304
19 P-T evolution of the carbonatitic melts and fluids
19 P-T evolution of the Carbonatitic Melts And fluids
Based on geobarometry studies on various mineral assemblages of the anorthosites of the KIC
(1,385-1,347 Ma) and the associated felsic suite (~1,380-1,340 Ma) pressures of 6-8 kbar have
been calculated for both their emplacement and subsolidus re-equilibration in the middle crust
(Drüppel et al., 2000; see Chapter 9). Following this, the emplacement depth of the younger
nepheline syenite and the ferrocarbonatites (1,140-1,120 Ma) is confined to pressures below 6
kbar. The lack of broad metasomatic aureoles surrounding the carbonatite dykes as well as the
sodic character of fenitization generally suggest the carbonatites to be deep-seated intrusions (Le
Bas, 1987, 1989). The P-T conditions for the magmatic crystallization and subsolidus reequilibration of the carbonatites and nepheline syenites and for the fenitization of anorthosites
and nepheline syenites were estimated from the chemical compositions and textural relationships
of their respective mineral assemblages, oxygen and sulphur isotope fractionation between
mineral pairs and fluid inclusion isochore data. The results of the various calculations are
summarized in the P-T plot of Fig. 19.4 (Drüppel et al., 2002a, subm.).
19.1 Emplacement and crystallization of the carbonatites
The following magmatic crystallization sequence was established for the carbonatitic breccia:
(1)
apatite, pyrochlore, biotite, ankerite, magnetite,
(2)
ankerite, interstitial calcite, magnetite,
(3)
chalcopyrite, pyrite, magnetite,
(4)
chalcopyrite, pyrite, millerite, polydymite/violarite, chalcocite/digenite, bornite,
magnetite, hematite,
(5)
hematite, chlorite, barite, strontianite, and calcite of the second generation.
In the cementing carbonatite matrix, apatite, pyrochlore and biotite crystallized first
accompanied and outlasted by ankerite and magnetite and followed by late interstitial calcite.
Dispersed grains of chalcopyrite and pyrite are most probably synchronous with the second
magnetite generation. During the end-stages of the carbonatite crystallization various sulphides,
hematite as well as late chlorite, barite, strontianite, and calcite of the second generation were
305
formed. The rocks suffered extensive and repeated deformation, which, however, mainly
affected ankerite and calcite.
To establish the P-T conditions during the crystallization of the carbonatites the following
approaches have been used: (1) biotite-apatite thermometry and HF barometry, (2) calcitedolomite thermometry, (3) equilibrium fractionation of oxygen isotopes in ankerite-magnetite
pairs, (4) equilibrium fractionation of sulphur isotopes in pyrite-chalcopyrite pairs and (5)
stability relations of Co-Ni-Fe sulphides using published thermodynamic data.
19.2.1 Biotite-apatite thermometry and HF barometry
The biotite-apatite geothermometer, first proposed by Stormer & Carmichael (1971),is based on
the partition coefficient of fluorine and hydroxyl between biotite and apatite. The
geothermometer was recalibrated by Ludington (1978), who additionally made allowance for the
effect of iron and magnesium substitution on fluorine in biotite. A set of consistent
thermodynamic properties of fluorine and chlorine end-member minerals and solid-solution
properties for biotite (considering reciprocal (Mg,Fe2+,AlVI) (F,Cl,OH) mixing properties) was
given by Zhu & Sverjensky (1991), who later proposed a formulation of the biotite-apatite
fluorine and hydroxyl geothermometer (Zhu & Sverjensky, 1992). The temperature-dependent FOH exchange between coexisting biotite and apatite is expressed by the reaction
0.5 KMg3AlSi3O10F2 + Ca5(PO4)3(OH) = 0.5 KMg3AlSi3O10(OH)2 + Ca5(PO4)3(F)
biotite
apatite
biotite
(19-1)
apatite
with the revised apatite-biotite geothermometer by Zhu & Sverjensky (1992) given by the
equation:
T (°C ) =
8852 − 0.024P(bars) + 5000 X Fe
1.987 ln K DAp, F/ Bt + 3.3666
where:
XFe
=
mole
(XFe=(Fe+AlVI)/(Fe+Mg+AlVI))
− 273.15
fraction
(19-2)
of
Fe
in
K DAp, F/ Bt =
(XF/XOH)Ap / (XF/XOH)Bt
XF, XOH =
mole fractions of F and OH in biotite and apatite
biotite
306
The pressure dependence of this geothermometer is negligible, whereas it is highly sensitive to
analytical errors in fluorine. The mean error (1σ) of the geothermometer is ± 50°C.
Sample
analyses
texture
Ku-98-131
6-2-Ap1
6-2-Bt1
Ku-99-SA8
KD
T (°C)
Pref: 3 kbar
T (°C)
Pref: 4 kbar
T (°C)
Pref: 5 kbar
log (ƒHF/ƒH2O)
core
core
rim
rim
rim
rim
12.29
11.65
15.14
16.88
36.93
31.71
1048
1065
985
955
774
804
1045
1062
982
952
771
802
1042
1059
980
950
769
800
-3.84
-3.82
-3.95
-4.00
-4.45
-4.36
6-2-Ap2
6-2-Bt4
core
core
rim
rim
10.20
10.38
15.97
33.15
1121
1115
981
805
1118
1112
979
803
1115
1109
976
800
-3.81
-3.82
-4.02
-4.42
6-2-Ap3
6-2-Bt2
core
rim
11.74
17.65
1074
954
1071
951
1068
949
-3.87
-4.07
6-2-Ap4
6-2-Bt2
core
rim
14.31
14.57
1013
1008
1010
1005
1007
1002
-3.97
-3.98
6-2-Ap5
6-2-Bt2
core
rim
8.96
21.21
1168
906
1165
904
1162
901
-3.75
-4.17
2-Ap1
2-Bt1
core
rim
9.91
13.19
1163
1067
1160
1064
1157
1061
-3.89
-4.01
2-Ap2
2-Bt2
core
core
rim
rim
10.68
11.64
19.17
13.11
1161
1132
981
1092
1158
1129
978
1090
1155
1126
976
1087
-3.90
-3.93
-4.15
-3.98
3-Ap1
3-Bt2
core
rim
9.69
12.42
1197
1110
1194
1107
1191
1104
-3.94
-4.04
4-Ap1
4-Bt1
core
rim
10.28
20.54
1156
946
1153
944
1150
941
-3.89
-4.20
Table. 19.1: Representative T and HF fugacity (log(ƒHF/ƒH2O) estimates for biotite-apatite pairs of the sodalite-rich
carbonatitic breccia, derived from biotite-apatite fluorine and hydroxyl geothermometry after Zhu & Sverjensky
(1992).
The calculation of the relative hydrogen fugacity followed the approach of Andersen &
Austrheim (1991), who suggested that the relative HF fugacity is independent of the phlogopite
composition, which contains negligible octahedral aluminium. The equation of Andersen &
Austrheim (1991), describing the relationship between apatite composition, temperature and the
relative HF fugacity of the coexisting fluid phase and/or volatile-containing magma, is expressed
as:
log (ƒHF/ƒH2O) = log(XF/XOH)Ap – 1.523 – 3200 / T
The uncertainty of the relative HF fugacity is c. ± 0.15 log units.
(19-3)
307
Temperatures have been calculated for individual biotite-apatite pairs in direct grain
contact from two samples of the sodalite-rich and two samples of the layered (sodalite-poor or –
free) carbonatitic breccia. The results obtained for temperatures and HF-fugacity data of the
sodalite-rich carbonatitic breccia and the layered carbonatitic breccia are compiled in Tables 19.1
and 19.2, respectively, and are illustrated in a log(fHF/fH2O) vs. 104/T diagram (Fig. 19.1).
Sample
analyses
texture
Ku-98-14
2-ap1/2-Bt1
Ku-98-48
KD
T (°C)
Pref: 3 kbar
T (°C)
Pref: 4 kbar
T (°C)
Pref: 5 kbar
log (fHF/fH2O)
core
100.56
647
645
643
-5.03
2-Ap2
2-Bt2
core
core
rim
rim
23.52
13.50
31.47
57.12
902
1054
835
720
900
1051
718
833
897
1048
716
831
-4.15
-3.88
-4.92
-4.05
2-Ap3
2-Bt2
core
core
rim
rim
19.46
14.01
39.07
71.03
950
1043
791
684
947
1040
788
682
945
1037
786
680
-4.06
-3.90
-4.43
-4.81
3-Ap1
2-Bt2
core
core
rim
recrystallised
recrystallised
12.85
14.16
58.66
163.66
147.90
1069
1039
716
566
583
1066
1036
714
565
581
1064
1034
712
563
579
-3.86
-3.90
-4.68
-5.42
-5.27
6-Ap1
2-Bt2
core
rim
44.51
45.95
765
759
763
757
761
755
-4.51
-4.53
a-Ap1
a-Bt4
core
core
rim
rim
rim
recrystallised
31.62
24.49
91.35
134.93
72.63
208.91
888
949
690
633
727
577
886
946
688
631
725
575
884
944
686
629
723
573
-4.49
-4.37
-5.07
-5.32
-4.93
-5.62
a-Ap2
a-Bt4
core
core
rim
rim
27.38
35.80
84.56
97.66
922
861
702
680
919
859
700
678
917
856
698
676
-4.42
-4.55
-5.02
-5.11
a-Ap3
a-Bt4
core
core
core
rim
rim
14.17
15.75
20.80
123.63
46.61
1103
1070
991
645
807
1100
1067
989
643
805
1097
1065
986
641
802
-4.13
-4.18
-4.30
-5.26
-4.69
Table. 19.2: Representative T and HF fugacity (log(ƒHF/ƒH2O) estimates for biotite-apatite pairs of the layered
carbonatitic breccia, derived from biotite-apatite fluorine and hydroxyl geothermometry after Zhu & Sverjensky
(1992).
Biotite-apatite pairs from the sodalite-rich carbonatitic breccia exhibit a cooling trend,
reaching continuously from near-liquidus temperatures of 1220°C (Wyllie, 1966) down to 770°C
at high corresponding relative HF fugacities (log(fHF/fH2O): -3.8 to –4.4). In contrast, samples of
the layered carbonatitic breccia indicate a lower temperature range (1070-630°C) at lower
relative HF fugacities (log(fHF/fH2O): -3.9 to -5.3). Recrystallised apatite-biotite pairs of the
308
layered carbonatitic breccia yield the lowest temperatures of 560-580°C and at the same time the
lowest relative HF fugacities of –5.3 to –5.6. Following this, it can be concluded that the original
F-OH distribution between apatite and biotite is not easily disturbed by secondary processes,
except recrystallisation, as has also been suggested by Andersen & Austrheim (1991) and Seifert
et al. (2000).
T (°C)
1200
-3.0
1000
800
700
600
-3.5
-4.5
-5.0
3
be
r
fo
rsi
te
A
an
d
al
nö
ite
SOLIDUS
LIQUIDUS
log (fHF/fH2O)
-4.0
RE-CRYSTALLISATION
1 g
limm
erite
and
2 g
limm
dolo
mitic
erite
lom
itic c
a
carb
arbo nd do
ona
nati
tite
te
sodalite-rich samples
Ku-98-131
Ku-99-SA8
-6.0
B
sodalite-poor samples
-5.5
Ku-98-14
Ku-98-14, recryst.
Ku-98-48
Ku-98-48, recryst.
6
7
8
9
10
11
12
10000/T (K)
Fig. 19.1: Log(ƒHF/ƒH2O) vs. 104/T diagram for magmatic and recrystallised apatite from the carbonatitic breccia of
the Swartbooisdrif alkaline suite. Arrow A: cooling trend derived from apatite-biotite pairs of the sodalite-rich
carbonatitic breccia, arrow B: cooling trend derived from apatite-biotite pairs of the layered (sodalite-poor)
carbonatitic breccia, arrows 1 and 2: cooling trends of glimmerite and dolomite carbonatite of the Delitzsch
complex, Germany (Seifert et al., 2000). arrow 3: cooling trend of alnöite of the Fen complex, Norway (Andersen &
Austrheim, 1991) and of berforsite of the Delitzsch complex, Germany (Seifert et al., 2000). The liquidus of T =
1200°C is derived from the maximum thermal stability of phlogopite (Yoder & Kushiro, 1969) whereas the solidus
of 625-650°C is suggested by the experimental data of Wyllie (1966).
309
The obtained trends for the two carbonatite subgroups are in excellent accordance with
the findings of Andersen & Austrheim (1991) and Seifert et al. (2000), who obtained similar
trends for their relative sequence of carbonatite intrusions in the Fen complex (Norway) and the
Delitzsch complex (Germany), respectively. The results indicate, that more than one magmatic
event must have produced the carbonatitic breccia. In addition it can be concluded, that the
transformation of albite into sodalite, which started during the end-stages and outlasted the
biotite formation in the sodalite-bearing carbonatitic breccia, occurred at relatively high
temperatures (up to ~770°C) early in the magmatic-metasomatic history.
19.2.2 Calcite-dolomite thermometry
Calcite-dolomite thermometry was applied using the calibration Anovitz & Essene (1987), who
provided the effect of Fe on the calcite-dolomite solvus. Temperatures were calculated for
calcite-ankerite pairs in direct grain contact from the carbonatitic breccia and the fenitized
syenite. The calculation yielded temperatures in the range of 410-530°C, the highest average
values of 470°C being obtained by the REE-rich carbonatitic breccia sample Ku-98-130a (Table
19.3). When compared to the results of the apatite-biotite thermometry (see Chapter 19.2.1) these
temperatures are relatively low, suggesting extensive recrystallisation and/or re-equilibration of
the carbonates from the carbonatitic breccia, which is in excellent accordance with the
petrographical findings.
Method of investigation
Rock type
Sample
Calcite-dolomite thermometry
CB
S, f
Textural relationship
Tmin
Tmax
Taverage
Ku-98-130b brownish carbonate aggregates
436°C
526°C
470°C
Ku-98-59b
406°C
534°C
438°C
Anovitz & Essene (1987)
matrix overgrowths
Table 19.3: Temperature estimates for calcite-ankerite pairs of the carbonatitic breccia (CB) and the fenitized
syenites (FS), derived from calcite-dolomite geothermometry after Anovitz & Essene (1987).
19.1.3 Ankerite-magnetite oxygen isotope equilibria
Stable isotope thermometry is an exchange thermometer, based on the partitioning of the heavy
(18O) and the light (16O) isotope between coexisting phases. The fractionation between two
mineral phases a and b (∆18O(a-b), generally expressed as 1000lnα(a-b)) varies as a function of
temperature, whereas pressure has no effect on the stable isotope fractionation. The temperature-
310
dependent fractionation between two phases (a, b) can be generally expressed as (e.g. Faure,
1986; Hoefs, 1997):
∆18O(a-b) = 1000lnα(a-b) = A·106·1/T2 + B
(19-4)
Since published data for oxygen isotopic equilibria between mineral pairs, such as ankeritemagnetite, are missing, the fractionation factors have to be derived from the respective mineralwater curves. The 1000lnα values for ankerite-water and magnetite-water were calculated
through the relations:
1000lnα(Ank-water) = 4.12·106·1/T2 - 4.62·103·1/T + 1.71
(19-4)
1000lnα(Mag-water) = 3.02·106·1/T2 – 12.00·103·1/T + 3.31
(19-5)
using the magnetite-water and the ankerite-water fractionation curves of Zheng (1991, 1999),
calibrated for temperature ranges of 0-1200°C. Temperatures for equilibrium fractionation
between ankerite and magnetite were estimated subsequently, by combining equation (19-6) and
(19-7).
Method of investigation
Equilibrium fractionation of
Rock type
Sample
δ18OSMOW (Mag)
δ18OSMOW (Ank)
T (°C)
CB
Ku-97-26
-1.82
8.95
449
CB
Ku-99-02
-4.21
9.07
344
CB
Ku-99-SA11
-1.46
9.73
479
FV
Ku-01-03
-2.63
8.91
413
oxygen isotopes in ankeritemagnetite pairs
(Zheng, 1991, 1999)
Table 19.4: Temperature estimates for ankerite-magnetite pairs of the carbonatitic breccia (CB) and the late
ferrocarbonatite veins (FV), calculated for ankerite-magnetite oxygen isotope equilibrium fractionation using
mineral-water fractionation curves of Zheng (1991, 1999).
The temperatures calculated for oxygen isotope equilibration of magnetite-ankerite pairs
from the carbonatitic breccia range between 344°C and 479°C (Table 19.4), thus being even
lower than those obtained by calcite-dolomite thermometry (440-530°C; see Chapter 19.2.2). A
similar temperature of 413°C was estimated for an ankerite-magnetite pair from a sample of a
late-stage ferrocarbonatite vein (Table 19.4). The lower temperatures obtained by the oxygen
311
isotopic ratios of ankerite-magnetite are interpreted in terms of a weak subsolidus reequilibration of the carbonatites, which is also supported by the oxygen isotope values of the
ankerites (see Chapter 18.1.3).
19.1.4 Pyrite-chalcopyrite sulphur isotope equilibria
The temperatures for the equilibrium fractionation of sulphur between coexisting pyrite and
chalcopyrite were estimated indirectly, by combining the respective mineral-H2S curves for
temperature ranges of 200-600°C and 200-700°C, respectively, of Ohmoto & Goldhaber (1997),
which are based on the general formulation (19-4):
1000lnα(Py-H2S) = 0.40·106·1/T2
(19-6)
1000lnα(Cpy-H2S) = -0.05·106·1/T2
(19-7)
Temperatures were calculated for three pairs of coexisting pyrite and chalcopyrite in latestage sulphide veins from a REE-rich sample of the carbonatitic breccia (Ku-98-130b). The
calculated temperatures range between 465°C and 488°C, thus suggesting that the sodalitization
of anorthositic xenoliths, incorporated by the carbonatitic breccia, is confined to temperatures >
490°C (Drüppel & Wagner, in prep.). The fact, that the T estimates for the sulphides, which
presumably formed during the end-stages of the carbonatite formation, are in the range of those
calculated for calcite-ankerite pairs (440-530°C) and even higher than those obtained for
ankerite-magnetite pairs (344-479°C) constrains the interpretation that at least the carbonates
were affected by secondary processes such as recrystallisation and/or re-equilibration, which,
however, did not effect the sulphur isotopic ratios of the investigated sulphides.
19.1.5 Stability relations of Co-Ni-Fe sulphides
The stability relations between various sulphide and oxide phases from samples of the
carbonatitic breccia coexisting with an S2-O2 fluid have been investigated in some detail by von
Seckendorff & Drüppel (1999) and von Seckendorff et al. (2000) and are illustrated in log ƒ(S2)
vs. 1000/T and log ƒ(S2) vs. log ƒ(O2) diagrams, both constructed for 1 bar total pressure (Fig.
19.2a, b; modified after von Seckendorff & Drüppel, 1999; von Seckendorff et al., 2000).
Stabilities of the different phases were calculated using published thermodynamic data (Toulmin
& Barton, 1964; Barton, 1973; Barton & Skinner, 1979). Since composition-activity
312
relationships are available for only some of the investigated phases, the constructed log ƒ(S2) vs.
1000/T diagram must be regarded as semi-quantitative. The findings of von Seckendorff &
Drüppel (1999) and von Seckendorff et al. (2000) for the crystallization sequence of the sulphide
and oxide phases are summarized below.
(1) The starting point of the reconstructed T- ƒ(S2) path (Fig. 19.2a) is given by coexisting
igneous pyrrhotite and included magnetite, with the ƒ(S2) and ƒ(O2) values being buffered by the
reaction (Fig. 19.2b)
6 FeS + 4O2 = 2 FeFe2O4 + 3S2
(19-8)
During, subsequent cooling, high-T pyrrhotite broke down to lamellar intergrowths of at least
two low-T pyrrhotite polytypes. A simultaneous increase of ƒ(S2) and ƒ(O2) (Fig. 19.2a, b), most
probably accompanied by cooling, led to a coupled replacement of these pyrrhotite intergrowths
by pyrite + hematite (± magnetite), according to reaction
4FeS + 2O2 = FeS2 + FeFe2O4 + S2
(19-9)
followed by the reaction
2FeFe2O4 + 2S2 = 2FeS2 + 2Fe2O3 + O2
(19-10)
The absence of vaesite suggests that the assemblage pyrite-hematite-magnetite became stable at
temperatures <320°C and log ƒ(S2) <-8 (Fig. 19.2a).
(2) No data are available to constrain the stability fields of siegenite, but from the established
stability fields of polydymite and linnaeite (Rosenqvist, 1954; Leegaard & Rosenqvist, 1964; cf.
Craig & Scott, 1974; Barton & Skinner, 1979), its maximum stability can be estimated at about
450°C with log ƒ(S2) in the vicinity of –6 (Fig. 19.2a). The replacement of siegenite by
chalcopyrite limits ƒ(S2) to values higher than the reaction:
iss + S2 = CuFeS2 + FeS2
(19-11)
(3) Millerite was probably derived from precursor (Ni,Fe)1-xS, stable at temperatures above
300°C (Fig. 19.2a). On the basis of the textural evidence, millerite coexisted with chalcopyrite
and pyrite. In the absence of vaesite, the narrow stability field of Ni1-xS + chalcopyrite + pyrite
extends from about 450°C and log ƒ(S2) = -3 to the curve NiS = Ni1-xS, where the stability field
of millerite + chalcopyrite + pyrite is reached at about 290°C and log ƒ(S2) = -10.5 (Fig. 19.2a).
(4) The replacement of millerite by polydymite-violarite took place with decreasing temperature.
Since aggregates of millerite, polydymite, chalcopyrite and pyrite are transsected or enclosed by
veins of magnetite and titanohematite, ƒ(S2) is most probably constrained by the buffer curve of
a)
313
2
2
0
e
curv
ion
t
a
s
nden
r co
u
f
l
su
m
y
-2
-6
Vs
-10
Pd 4 Mil
t
Ca
3
Ni1- S
x
Ln 8 NiS
S
Co 9
300
2.0
1.8
5
2
y
+P
s
is at
C
Py
Po
0
-2
1
-4
S
-x
2
Ni 1 i 3+xS
N
Ln
S4
o3 S
C o 1-x
C
-6
-8
S
x
Ni 1- S 6
Ni 7
Ln
NiS 6
S
Ni 7
-14
-16
Py
p+ s
c
s
C i
S
-x
Ni 1
-8
-12
e
+H
Py Mag
+P
Bn cp
C
C
i s cp
s
log f(S2)
-4
Py
is s + p
Cc
Vs S
Ni 1-x
-10
Po
Fe
-12
-14
T [°C]
500
400
1.6
1.4
600
700 800 900
1.2
1.0
0.8
-16
1000/T [1/K]
b)
0
log f(S2)
-5
sulfur condensation curve
Vs
-10
Mil
-15
Ni7S6
Pd Cat
Ln
Co9S8
Bn+Py
Ccp
Ccp+Py
Py
FeSO4
iss
Po
-20
-25
-30
-35
Mag
Fe
-40
-45
-50
-100
Hem
NNO
325°C, 1 bar
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
log f(O2)
Fig. 19.2: a) Diagram log ƒ(S2) vs. 1000/T (modified after von Seckendorff & Drüppel, 1999; von Seckendorff et
al., 2000) using thermodynamic data by Toulmin & Barton (1964), Barton (1973) and Barton & Skinner (1979). The
stability fields for polydymite and linnaeite are shown in grey. The arrow depicts the reconstructed T- ƒ(S2) path. b)
Isothermal section of log ƒ(S2) vs. log ƒ(O2) at 325°C, 1 bar, calculated from thermodynamic data of Barton &
Skinner (1979), following the approach of Holland (1959). For further details see text (Bn bornite, Cat cattierite,
Ccp chalcopyrite, Fe metallic iron, Hem hematite, iss intermediate solid solution, Ln linnaeite, Mag magnetite, Mil
millerite, NNO Ni-NiO-buffer, Pd polydymite, Po pyrrhotite, Py pyrite, Vs vaesite).
314
reaction (19-10), that is positioned within the polydymite stability field, thus suggesting that the
replacement did not go to completion.
(5) The pyrites occurring in late-stage sulphide veins, which transsect the oxide-sulphide
aggregates appear to be the latest sulphides precipitated. As was assumed by von Seckendorff &
Drüppel (1999) and von Seckendorff et al. (2000), this process took place during late
hydrothermal alteration of the rocks, outlasting the metasomatic event. However, the high
temperatures obtained for coexisting pyrite and chalcopyrite (± magnetite), occurring in late
sulphide-oxide veins (465 - 488°C; see Chapter 19.1.4), constrain the log ƒ(S2) at comparably
high values of c. –6, thus suggesting that the sulphide veins were rather formed during the
repeated intrusion of carbonatite magmas than by low-temperature hydrothermal alteration of the
To summarize, the semi-quantitative T-ƒ(S2) path of von Seckendorff & Drüppel (1999)
and von Seckendorff et al. (2000; Fig. 19.2a) starts with igneous, high temperature conditions (c.
700-800°C), documented by coexisting ilmenite, Ti-magnetite and pyrrhotite, crossed the
stability fields of iss, Ni1-xS and siegenite, and ended in the polydymite stability field close to the
magnetite + S2 = pyrite + hematite buffer curve at <280°C and log ƒ(S2) <-11. The formation of
late sulphide veins at comparably high temperatures of 465-488°C and log ƒ(S2) of c.–6 are
ascribed to repeated magmatic activity, leading to a renewed elevation of both the temperature
and the ƒ(S2).
19.2 Emplacement and fenitization of the nepheline syenite
Temperatures of the nepheline syenite crystallization can be estimated from the chemical
composition of orthomagmatic, concentrically zoned nepheline, using the nepheline liquidus
thermometer after Hamilton (1964), which provides a temperature scale for the varying amounts
of excess SiO2 in natural nepheline coexisting with K-feldspar. The position of the approximate
limit of nepheline solid-solution at 1068°C (Fig. 19.3) is based on determinations of the
nepheline solid-solution limits in the systems NaAlSiO4-NaAlSi3O8 (Greig & Barth, 1938) and
NaAlSiO4-KAlSiO4 (Tuttle & Smith, 1958).
Calculated temperatures reveal a large spread, reaching from above the 1068°C isotherm
for the SiO2-rich core compositions to temperatures as low as ~770°C at the SiO2-poor rims of
315
the individual nepheline grains (Fig. 19.3). In clear contrast, the outermost Ne-rich margins of
the nepheline crystals, that were presumably formed by the interaction between the nephelinesyenite and the Na-rich and Si-deficient carbonatite magma, indicate lower temperatures of 700775°C and may thus give a first idea of the temperature range of the wall-rock fenitization. Since
both, the analyses of the nepheline mineral chemistry and the T estimates point to continuous
nepheline growth, it might be possible that Na was introduced to the system before nepheline
was fully crystallized.
SiO2
80
1068°
C
775
°C
90
nepheline syenite
Ku-99-15
core
rim
outermost margin, cracks
core
700°C
500°C
100
NaAlSiO4
90
rim
80
70
60
KAlSiO4
Fig. 19.3: Nepheline liquidus thermometry of nepheline from the Swartbooisdrif nepheline syenites after Hamilton
(1964).
Moreover, the temperatures of the sodalitization of nepheline can be estimated from the
stability field of sodalite formed by reaction (12-1), in relation to the NaCl concentration of the
interacting fluids and temperature, based on the experimental investigations of Kostelńikov &
Zhornyak (1995; see Chapter 12.1). Applying plausible ranges of 19-30 wt.% NaCl eq. for the
primary salt concentrations of the fenitizing fluids (see Chapter 17.1), corresponding maximum
temperatures of c. 630-730°C are obtained for the transformation of nepheline into sodalite, good
in accordance with the temperatures obtained by Ne liquidus thermometry. Applying these Testimates to the experimentally derived P-T stability fields of sodalite in the system NaAlSiO4SiO2-NaCl of Sharp et al. (1989) and using reaction (12-1): 6 Ne + 2 NaCl → Sdl (see Chapter
12.2), the sodalite stability with respect to nepheline is confined to pressures < 8 kbar.
316
19.3 Fenitization of anorthositic xenoliths and syenites– implications from fluidinclusion isochore data
In order to constrain the P-T ranges of (1) the sodalite-formation in anorthosite xenoliths,
incorporated by the carbonatitic breccia, and (2) the fenitization of syenites bordering the
carbonatite centres, petrographical findings as well as temperature calculations for the
carbonatite minerals were combined with fluid-inclusion isochore data. The resulting P-T paths
are illustrated in Fig. 19.4a and Fig. 19.4b, respectively.
For an intermediate temperature range of 410-530°C, obtained for the carbonate
recrystallisation (see Chapter 19.2.2), corresponding pressures of c. 4-5 kbar were obtained,
using isochores calculated for pseudosecondary aqueous inclusions in ankerite from the REErich carbonatitic breccia (Fig. 19.4a). The resulting temperatures of sodalite formation for this P
range are considerably variable, reaching from 800°C down to 530°C, using isochores of
aqueous inclusions hosted by sodalite from the two REE-rich samples of the carbonatitic breccia
(Ku-98-130a, Ku-98-130b). A more restricted T-range (725-625°C) is obtained for isochores of
aqueous fluid inclusions in sodalite from a sample of the sodalite-rich carbonatitic breccia.
Regarding the effects of secondary processes responsible for the REE-enrichment of the REErich carbonatitic breccia, the temperatures obtained for the REE-poor sample Ku-98-131 are
suggested to give the more likely T conditions for transformation of albite into sodalite. It has to
be mentioned, however, that the obtained temperature range for the sodalite formation is in
excellent accordance with the previous findings, that (1) sodalite formation is confined to
temperatures below those obtained for the co-crystallization of apatite and biotite (i.e. below c.
770°C) and higher than those calculated for coexisting pyrite and chalcopyrite (i.e. 465-488°C)
and (2) sodalitization of nepheline of the bordering nepheline syenites presumably occurred at
similar temperatures (c. 700-775°C). It is also noteworthy, that the formation of large masses of
sodalite, both in the bordering anorthosites and in the fenitized anorthosite xenoliths can not be
explained by a short magmatic episode in a narrow restricted temperature range but needs large
amounts of NaCl-rich fluids, expelled during a certain time interval and temperature range.
For a given pressure range of c. 4-5 kbar, that is assumed to reflect the emplacement
depth of the carbonatitic breccia (c. 12-15 km), corresponding temperatures of 330-500°C were
obtained for the fenitization of the syenites, using isochores calculated for secondary fluid
317
6
Formation of
sodalite
5
pressure (kbar)
Crystallisation
of apatite and biotite
4
Recrystallisation
of carbonate
A&E
3
A&E
Z&S
Z&S
2
aqueous inclusions
Sdl (Ku-98-130a)
Sdl (Ku-98-130b)
Sdl (Ku-98-131)
1
Ank (Ku-98-130b)
a
0
0
100 200 300 400 500 600 700 800 900 1000 1100 1200
temperature (°C)
6
CO2-inclusions
Qtz
(Ku-98-40)
5
Formation of
sodalite
pressure (kbar)
Crystallisation
of apatite and biotite
4
Fenitization
of syenites
Z&S
3
Z&S
2
aqueous inclusions
hydrothermal
alteration of
syenites
Sdl (Ku-98-130a)
Sdl (Ku-98-130b)
Sdl (Ku-98-131)
1
Qtz (Ku-98-40)
b
0
0
100 200 300 400 500 600 700 800 900 1000 1100 1200
temperature (°C)
Fig. 19.4: P-T diagram for the carbonatite emplacement and the carbonatite-induced fenitizing processes in the
Swartbooisdrif area, as deduced by conventional geothermometry combined with fluid-inclusion isochore data (see
text for more detail). a) Carbonatite (re-)crystallization and fenitization of anorthositic xenoliths incorporated by the
carbonatitic breccia. b) Fenitization and low-temperature alteration of syenites. (Arrows in a) and b) mark the
respective evolutionary trends of the investigated rock units; A&E calcite-dolomite geothermometry after Anovitz &
Essene, 1987, Z&S biotite-apatite fluorine and hydroxyl geothermometer after Zhu & Sverjensky, 1992).
318
inclusions in quartz of a quartz-syenite sample (Ku-98-40; Fig. 19.4b). Even though this sample
suffered only minor alteration by the fenitizing solutions (e.g. a weak alteration of K-feldspar to
albite), the obtained T-range is suggested to be a good proxy for the fenitization of the bordering
syenites, since (1) steep temperature gradients in the vicinity of the carbonatitic centres can be
assumed, judging from the lack of broad metasomatic aureoles surrounding the carbonatitic
breccia, (2) the replacement of K-feldspar by albite-muscovite assemblages does not need high
temperatures and (3) sodalite, which presumably formed under higher temperatures of 530800°C, is virtually completely absent in samples of the fenitized syenites. Moreover, the low Trange obtained for the fenitization of the lithologies bordering the carbonatites, may also account
for the formation of comparably narrow metasomatic aureoles in the anorthositic wall-rocks,
where sodalite is restricted to the direct contacts between the carbonatite dykes and the
anorthosites and hence confined to zones that rarely reach a width of 60 cm.
As textural evidence for a cogenetic formation of the CO2 and the H2O inclusions in
quartz of the quartz-syenite is lacking, the late CO2 inclusion trails are ascribed to a weak low
temperature hydrothermal overprint of the syenites. Assuming a plausible T-range for
hydrothermal activity (c. 100-300°C), corresponding pressures range of < 0.5 kbar to 1.5 kbar, in
good accordance with an alteration by circulating meteoric waters confined to the shear-zones.
To summarize, the emplacement of the carbonatitic breccia at a depth of c. 12-15 km (c.
4-5 kbar) was followed by the early crystallization of apatite and biotite (± ankerite and
magnetite) under igneous, high-temperature conditions (1200-770°C). Subsequently, NaCl-rich
carbonatite fluids caused the partial replacement of albite-rich anorthosite xenoliths by sodalite
under still high temperatures of 770-530°C. In the subsolidus stage the carbonates
(ankerite/calcite) underwent extensive recrystallisation, confined to temperatures of c. 410530°C. The fenitization of the syenites bordering the carbonatite centres occurred under
comparably low temperatures of 330-500°C. The low-temperature alteration of the syenites by
CO2-rich solutions presumably reflects an independent event, restricted to subvolcanic crustal
levels (<0.5 kbar to 1.5 kbar) and thus post-dating the magmatic and metasomatic evolution of
the carbonatites.
20 Conclusions
319
20 Conclusions
20.1 Origin and evolution of the Swartbooisdrif carbonatites
There is striking field evidence that both the carbonatitic breccia and the late stage
ferrocarbonatite veins are magmatic rather than hydrothermal in origin. They occur as dykes and
veins with cross-cutting relationships and margins disturbed by fenitic aureoles, and contain
abundant flow-oriented xenoliths from the surrounding anorthosite and syenite. In addition, the
mineral assemblage of both carbonatite generations
•
ankerite + calcite + magnetite + biotite ± apatite ± pyrochlore in the REE-poor
carbonatitic breccia and
•
ankerite + calcite + magnetite ± pyrochlore ± rutile in the late ferrocarbonatite veins,
their geochemical characteristics (high contents of incompatible elements like Sr, Ba and REE)
and the values of O and C isotope ratios (8.91 to 9.73 and –6.73 to –6.98, respectively) again
indicate igneous derivation, with the δ18O values suggesting only minor subsolidus alteration.
Following these arguments the carbonatites of both generations can be described as mantlederived, magmatic ferrocarbonatite magmas.
At least two magmatic events must have produced the carbonatitic breccia and the late
ferrocarbonatite veins as is evidenced by (i) their temporal relationship (ii) their differing mineral
assemblages and (iii) differences in their geochemical signatures. Moreover, evidence is
provided that the carbonatitic breccia itself was formed by multiple magmatic events, i.e.: (1) the
abundance of clasts of carbonatitic breccia cemented by a carbonate matrix, (2) elemental
recurrences in the zoning patterns of apatite, biotite and ankerite, (3) repeated emanation of high
salinity fluids recorded by quartz of the bordering quartz syenites, (4) the formation of texturally
late pyrite under temperatures, which are higher than those obtained for previously crystallized
sulphides and (5) the observation of at least two different cooling trends for sodalite-bearing and
sodalite-free samples of the carbonatitic breccia, determined by apatite-biotite thermometry and
HF barometry.
Based on general differences in their magmatic mineralogy and their mode of
fenitization, the following simplified evolutionary trend can be established for the succession of
the various carbonatite intrusions:
320
20 Conclusions
1. Magmatic evolution of the carbonatitic breccia and the ferrocarbonatite veins
The local presence of apatite and biotite phenocrysts in the carbonatitic breccia is suggestive of a
ferrocarbonatite melt that additionally contained minor amounts of P, K, Si, and Al. The absence
of these phases in the ferrocarbonatite veins, which are depleted in Fe, Mg, Si, Al, K, Ti, P, and
S with respect to the carbonatitic breccia, implies that the ferrocarbonatite veins represent latestage fractionation products of the ferrocarbonatite magma which forms the carbonatitic breccia,
with their emplacement post-dating the extensive crystallization of biotite, ankerite, Fe-Ti
oxides, apatite, and sulphides. However, the carbonatite-related fenitization style (sodic, hydrous
fenitization), the primary composition of the fenitizing solution (Sr- and REE-bearing aqueous
brine, virtually free of CO2 and Ca), the trace-element composition of apatite (high SREE, high
(La/Yb)cn and (La/Nd)cn ratios) and the chondrite-normalized trace element plots of the
carbonatitic breccia (troughs at K, Sr, P, Ti), suggest that fractionation was even active previous
to the emplacement of the older carbonatite unit, the carbonatitic breccia.
The high modal amounts of metasomatically formed sodalite in most samples of the
carbonatitic breccia point to a high metasomatic potential of this rock type and, moreover,
suggest, that these melts contained appreciable amounts of sodium before alkali-loss during
fenitization. The abundance of sodalite correlates with high temperatures of apatite-biotite
equilibria (~1200-770°C). In contrast, sodalite-free samples of the carbonatitic breccia reach
lower apatite-biotite temperatures (~1070-630°C), thus evidencing that most of the fluids were
expelled during the early and high-T intrusion stages of this rock type. This interpretation is in
excellent agreement with the lack of extensive metasomatic aureoles around the youngest
recognisable carbonatite unit, the late-stage ferrocarbonatite veins, which incorporated
previously fenitized xenoliths but rarely show any wallrock alteration related to their
emplacement. The occurrence of late barite in samples of the carbonatitic breccia as well as the
sulphur isotopic data for pyrite and chalcopyrite imply a continuous trend towards higher oxygen
fugacities with carbonatite evolution.
2. REE-enrichment of the carbonatitic breccia
The REE mineralization of REE-rich samples of the carbonatitic breccia appears to be of
secondary origin, since the REE-minerals occur along the grain margins of ankerite, which
display a similar or even REE-depleted composition when compared to those of the
corresponding REE-poor samples. Moreover, the REE mineralization in the investigated samples
is accompanied by the additional occurrence of barite and strontianite, which are commonly
20 Conclusions
321
found in carbonatites exposed to secondary alteration. A supergene origin for the REE
enrichment can be ruled out, as ankerite from these rock types lacks fractionation to heavier
δ18O, but has 18O values in the range of ankerite from the REE-poor carbonatitic breccia. An in
situ mechanism of hydrothermal solution and re-precipitation is not likely, as a concentration
factor of over 100 times is required to achieve the REE-content of the REE-rich samples. Hence,
an injection of Sr and REE-enriched, late-stage carbonatite fluids must be considered. A process
like this would be in excellent accordance with the previous findings, i.e. (1) the REEenrichment recorded by the samples is not accompanied by simultaneous Na enrichment, hence
suggesting that the sodium contained in early fluids was already expelled during previous albite
and sodalite formation and (2) late-stage, aqueous fluid-inclusions hosted by recrystallised
ankerite, were found to be strongly enriched in Sr and the LREE, whereas they display
comparably low salt concentrations. Following these arguments the REE enrichment of the two
samples of the carbonatitic breccia resulted from a hydrothermal overprint by late-stage,
fractionated carbonatite fluids.
3. P-T conditions of the carbonatite emplacement and crystallization
Judging from the lack of broad metasomatic aureoles and the sodic style of fenitization the
Swartbooisdrif carbonatites may be generally assumed to be deep-seated intrusions.
Conformably, the pressures obtained for the (re-)crystallization of the carbonatites (c. 4-5 kbar)
point to emplacement depths of 12-15 km. The first minerals to crystallize are apatite, biotite,
magnetite, and ankerite. Temperatures obtained for apatite-biotite pairs from the carbonatitic
breccia reach from near-liquidus conditions of c. 1200°C (Wyllie, 1966) and proceed down to
near-solidus conditions of 630°C. The formation of pyrrhotite, coexisting with magnetite,
presumably occurred under high T conditions of c. 700-800°C. In contrast, most of the other
sulphides (pyrite, chalcopyrite, millerite, siegenite, polydymite-violarite, fletcherite) were
formed in the subsolidus stage, under comparably low temperatures of 490°C to < 280°C. Both
biotite and apatite as well as the carbonates were subjected to extensive recrystallisation under T
conditions of 560-580°C and 410-530°C, respectively.
4. The carbonatite-nepheline syenite association
The intimate association of the Swartbooisdrif carbonatites with nepheline-syenite and syenite
dykes might suggest a genetic relationship between these rock-types. However, the syenites are
~150 Ma older than the carbonatite intrusions, thus providing evidence that their close spatial
322
20 Conclusions
association rather results from the preferred ascent of both the older silicate and the younger
carbonatite magmas along pre-existing faults. It has to be mentioned, however, that no age has
yet been constrained for the Swartbooisdrif nepheline syenite, which differs significantly from
the ~1,380-1,340 Ma syenite dykes of the same area in both its mineralogy and its major and
trace element composition. It is hence possible, that the nepheline syenite and the
ferrocarbonatite are part of one single magmatic suite, even though further age determinations
are needed to constrain this assumption.
20.2 The breakdown of igneous minerals during fenitization
Fluids are commonly associated with carbonatite melts and are responsible for the formation of
fenitic aureoles in the bordering lithologies. The fluids expelled by the carbonatite melts of
Swartbooisdrif are mainly characterized by high NaCl- and H2O-amounts, as is evident from the
hydrous and sodalite-rich nature of the fenites. The composition and evolutionary trends of these
fenitizing agents were estimated from both the sequence of metasomatic reactions within
wallrock xenoliths in the carbonatitic breccia and fluid inclusions data (microthermometric fluid
inclusion measurements combined with synchrotron micro-XRF analyses):
(1) From textural evidence, it is clear that the formation of almost pure albite at the
expense of former plagioclase and alkali-feldspar from anorthosite and syenite, respectively,
predated the emplacement of the carbonatite melts. This process presumably occurred
simultaneously with the transformation of anhydrous Fe-Mg-silicates and amphibole into biotite
in the respective rock types. From this it can be concluded that the carbonatite magma yielded
water saturation already during its uprise and expelled large amounts of fluid, which migrated
along the pre-existing pathways and pervasively altered the brecciated anorthosite and syenite
exposed in the shear zones as well as the neighbouring lithologies. The metasomatic reactions
require that Na and H2O were the dominant components of this fluid, which, in addition,
presumably contained Fe, Cl and K. In contrast, the Si necessary for the transformation of Anrich plagioclase into almost pure albite is supposed to be derived from the fenites themselves
during Si-releasing metasomatic reactions, rather than being a primary constituent of the
fenitizing agents.
(2) After its final emplacement in the crust (~12-15 km), the carbonatite magma again
reached water saturation following cooling from ~1200°C to 770°C and fractionation of
anhydrous minerals (apatite, ankerite and magnetite). In a temperature range of 800-530°C
(maximum at 725-625°C) the circulation of NaCl-rich aqueous brines (20-30 wt.% NaCl eq.),
20 Conclusions
323
expelled by the crystallising carbonatite magma, caused the partial transformation of albite of the
incorporated anorthositic xenoliths into sodalite. The sodalitization of nepheline from a
nepheline syenite dyke in direct contact to the carbonatitic breccia, occurred under similar
temperatures of c. 730-630°C. This fluid, whose composition was presumably close to the initial
composition of the carbonatite fluid, was also capable of transporting minor amounts of Sr, Ba,
Fe, Nb and the LREE. Most of the Na contained in the fenitizing agents was consumed during
this major sodalite-forming event, which was restricted to the carbonatitic breccia and narrow
reaction-zones (up to 60 cm in width) within the bordering anorthosite, syenite and nepheline
syenite. Conformably, fluids trapped in quartz of a neighbouring quartz-syenite dyke display
highly variable, but generally lower salinities of 20-24 wt.% NaCl eq. and 9-17 wt.% NaCl eq.
and, at the same time, higher amounts of K, Ca, Fe, Ti, Sr and Ba, evidencing previous Na-loss
during extensive sodalite and/or albite formation. Moreover, the temperatures obtained for the
fenitization of the syenites (330-500°C) are distinctly lower than those of the direct contact zones
between the nepheline syenite and the carbonatitic breccia. The low temperatures of fenitization,
corresponding to comparably low salinities of the fenitizing fluids might be the reasons for the
general lack of sodalite in the fenitized syenite.
(3) The fenitic mineral assemblage of cancrinite and muscovite, replacing albite and postdating sodalite, implies that the evolved fluids were still aqueous brines but no longer contained
sufficient Na to form sodalite. The Si-free composition of the subsequently formed Na-Al phase
and its high Al:Na ratio of 4:1 when compared to sodalite (Al/Na < 1) strongly suggests that
even the late-stage carbonatite fluids were still strongly deficient in Si but contained only minor
amounts of Na. The K and Ca, required by most of the late metasomatic reactions, may be
primary constituents of these fluids or derived by K- and Ca-releasing reactions. In the case of
nepheline syenite, the close association of sodalite and muscovite, both formed at the expense of
nepheline, favours the latter model. On the other hand, the clear temporal succession of the
respective minerals observed in the other fenite types rather argues against a coupled process.
20.3 Links to other carbonatite complexes of NW Namibia and SW Angola
The ferrocarbonatite dykes of Swartbooisdrif have traditionally been grouped together with E-W
and NE-SW trending calcio-carbonatite, lamprophyre and Ne-syenite dykes of the Epembe area
(Ferguson et al., 1975; Menge, 1986, 1996; Verwoerd, 1993; von Seckendorff et al., 2000). In
addition, syenite, calciocarbonatite and magnesiocarbonatite dykes of the Proterozoic Lupongola
Complex of Angola, ~20 km north-west of Swartbooisdrif, have been considered part of the
324
20 Conclusions
same alkaline suite (Menge, 1996; Alberti et al., 2000; Thompson et al., 2002). Support for a
genetic relation between at least the Lupongola and the Epembe suite is given by rock samples
collected south of the Namibian-Angolan border about 2 km distant from the Lupongola
Complex and forming the north-eastern extension of the NE trending Epembe dykes. In this area,
the dykes are almost entirely composed of calciocarbonatites and magnesiocarbonatites with
minor syenites, similar in mineralogy and geochemistry to those observed in the Epembe
(Ferguson et al., 1975) and the Lupongola alkaline suite (Alberti et al., 2000). In clear contrast to
the Swartbooisdrif ferrocarbonatites, however, the carbonatites in these areas lack the extensive
formation of sodalite-bearing metasomatic aureoles and xenoliths.
Based on Sm-Nd and Rb-Sr isotopic data Alberti et al. (2000) concluded, that the age of
the Lupongola carbonatites, even though yet undated, could be ~1,100 Ma. However, U-Pb
pyrochlore ages of ~1,204 Ma have recently been constrained for the Epembe carbonatites
whereas the Swartbooisdrif carbonatites yielded a distinctly younger U-Pb-pyrochlore age of
~1,140-1,120 Ma (Littmann et al., in prep.). These age constraints strongly suggest that at least
the Epembe and the Swartbooisdrif carbonatitic centres represent independent intrusions.
20.4 Concluding remarks
The present study has shown that the generation of carbonatite magmas and the
associated contact metasomatism in the Swartbooisdrif area are the result of several magmatic
and interrelated metasomatic processes, each inducing characteristic mineral reactions in the
anorthositic and syenitic wall-rock. It is possible to relate the metasomatic reactions to both, the
temporal succession of carbonatite intrusions and their respective magmatic crystallization
history.
The metasomatic processes related to carbonatites can convey important information on
the nature and evolution of the carbonatite melts. All present findings suggest that the
Swartbooisdrif ferrocarbonatite intrusions represent deeply emplaced and highly evolved
magmatic carbonatite melts, which contained appreciable amounts of sodium before alkali-loss
during fenitization.
21 References
325
21 References
Alberti, A., Castorina, F., Censi, P., Comin-Chiaramonti, P. & Gomes, C.B. (2000):
Geochemical characteristics of Cretaceous carbonatites from Angola. Journal of African
Earth Sciences 29: 735-759.
Andersen, T. (1984): Secondary processes in carbonatites: Petrology of "rødberg" (hematitecalcite-dolomite carbonatite) in the Fen complex, Telemark (South Norway). Lithos 17:
227-245.
Andersen, T. (1986): Magmatic fluids in the Fen carbonatite complex, S.E. Norway: Evidence
for mid-crustal fractionation from solid and fluid inclusions in apatite. Contributions to
Mineralogy and Petrology 93: 491-503.
Andersen, T. (1987): Mantle and crustal components in a carbonatite complex, and the evolution
of carbonatite magma: REE and isotopic evidence from the Fen Complex, southeast
Norway. Chemical Geology 65: 147-166.
Andersen, T. (1989): Carbonatite-related contact metasomatism in the Fen complex, Norway:
effects and petrogenetic implications. Mineralogical Magazine 53: 395-414.
Andersen, T. & Austrheim, H. (1991): Temperature-HF fugacity trends during crystallisation of
calcite carbonatite magma in the Fen complex, Norway. Mineralogical Magazine 55: 8194.
Anderson, J.L. (1980): Mineral equilibria and crystallization conditions in the late Precambrian
Wolf River rapakivi massif, Wisconsin. American Journal of Science 280: 289-332.
Anderson, J.L. (1983): Proterzoic anorogenic granite plutonism of North America. In: Medaris,
L.G., Byers, C.W., Mickelson, D.M. & Shanks, W.C. (eds.): Proterozoic geology: selected
papers from an International Proterozoic Symposium. Geological Society of America,
Special Papers 161, pp. 133-154.
Anderson, J.L. & Bender, E.E. (1989): Nature and origin of Proterozoic A-type granitic
magmatism in the southwestern United States of America. Lithos 23: 19-52.
Anderson, A.T. & Morin, M. (1968): Two types of massif anorthosite and their implications
regarding the thermal history of the crust. In: Isachsen Y.W. (ed.): Origin of anorthosite
and related rocks. N.Y. State Museum and Science Service, Memoir 18, pp. 47-55.
Andrade, F.R.D., Möller, P. & Höhndorf, A. (1999): The effect of hydrothermal alteration on
the Sr and Nd isotopic signatures of the Barro do Itapirapuã carbonatite, southern Brazil.
The Journal of Geology 107: 177-191.
326
21 References
Anovitz, L.M. & Essene, E.J. (1987): Phase equilibria in the system CaCO3-MgCO3-FeCO3.
Journal of Petrology 28: 389-414.
Ashwal, L.D. (1993): Anorthosites. Minerals and Rocks Series 21. Springer-Verlag, Berlin, 422
pp.
Ashwal, L.D., Hamilton, M.A., Morel, V.P.I. & Rambeloson, R.A. (1998): Geology, petrology
and isotope geochemistry of massif-type anorthosites from southwest Madagascar.
Contributions to Mineralogy and Petrology 133: 389-401.
Ashwal, L.D. & Twist, D. (1994): The Kunene complex, Angola/Namibia: a composite massiftype anorthosite complex. Geological Magazine 131: 579-591.
Barker, D.S. (1976): Phase relations in the system NaAlSiO4-SiO2-NaCl-H2O at 400°-800°C and
1 kilobar, and petrologic implications. The Journal of Geology 84: 97-106.
Barton, M., Sheets, J.M., Lee, W.E. & van Gaans, C. (1991): Occurrence of low-Ca
clinopyroxene and the role of deformation in the formation of pyroxene - Fe-Ti oxide
symplectites. Contributions to Mineralogy and Petrology 108: 181-195.
Barton, M. & van Gaans, C. (1988): Formation of orthopyroxene-Fe-Ti oxide symplectites in
Precambrian intrusives, Rogaland, southwestern Norway. American Mineralogist 73:
1046-1059.
Barton, P.B. Jr. (1973): Solid solutions in system Cu-Fe-S Part I. The Cu-S and Cu-Fe-S join.
Economic Geology 68: 455-465.
Barton, P.B. Jr. & Skinner, B.J. (1979): Sulphide mineral stabilities. In: Barnes, H.L. (ed.):
Geochemistry of hydrothermal ore deposits. Holt, Rinehart and Winston, New York, pp.
236-333.
Beard, A.D., Downes, H., Vetrin, V., Kempton, P.D. & Maluski, H. (1996): Petrogenesis of
Devonian lamprophyre and carbonatite minor intrusions, Kandalaksha Gulf (Kola
Peninsula, Russia). Lithos 39: 93-119
Beetz P.F.W. (1933): Geology of South West Angola, between Cunene and Lunda Axis.
Transactions of the Geological Society of South Africa 36: 137-176.
Bell, K. & Keller, J. (eds.) (1995): Carbonatite volcanism: Oldoinyo Lengai and the petrogenesis
of natrocarbonatites. Springer Verlag, Berlin, 210 pp.
Berg, J.H. (1977a): Dry granulite mineral assemblages in the contact aureoles of the Nain
Complex, Labrador. Contributions to Mineralogy and Petrology 64: 33-52.
Berg, J.H. (1977b): Regional geobarometry in the contact aureoles of the anorthositic Nain
Complex, Labrador. Journal of Petrology 18: 399-430.
21 References
327
Berg, L.-P., Shamsher, M.K., El-Daher, S.S., Kakkar, V.V., Authi, K.S., Buchanan, P.C.,
Zolensky, M.E. & Reid, A.M. (1997): Petrology of Allende dark inclusions. Geochimica et
Cosmochimica Acta 61: 1733-1743.
Bhattacharya, A., Krishnakumar, K.R., Raith, M. & Sen, S.K. (1991): An improved set of a-X
parameters for Fe-Mg-Ca garnets and refinements of the orthopyroxene-garnet
thermometer and the orthopyroxene-garnet-plagioclase-quartz barometer. Journal of
Petrology 32: 629-656.
Bhattacharya, A., Mohanty, L., Maji, A., Sen, S.K. & Raith, M. (1992): Non-ideal mixing in the
phlogopite-annite binary: constraints from experimental data on Mg-Fe partitioning and a
reformulation of the biotite-garnet geothermometer. Contributions to Mineralogy and
Bingen, B., Demaiffe, D. & Delhal, J. (1984): Petrologic and geothermobarometric
investigations in the Kasai gabbro-noritic complex and associated metadolerite dykes
(Zaire). Bulletin de Minéralogie 107: 665-682.
Binstedt, N. (1981): The stability of sodalite minerals. Progress in experimental petrology,
N.E.R.C. Report 5: 54-56.
Bodnar, J.R. (1993): Revised equation and table for determining the freezing point depression of
H2O-NaCl solutions. Geochimica et Cosmochimica Acta 57: 683-684.
Bohlen, S.R., Wall, V.J. & Boettcher, A.L. (1983): Geobarometry in granulites. In: Saxena, S.K.
(ed.): Kinetics and equilibrium in mineral reactions. Advances in physical geochemistry,
Volume 3, Springer-Verlag, New York, pp. 141-172.
Bonshtedt-Kupletskaya, E.M. (1966): On the questions of systematics of the pyrochloremicrolite group. (in Russian). Zapiski Vsesoyuznogo Mineralogicheskogo Obshchestva 95:
134-144.
Borthwick, J. & Harmon, R.S. (1982): A note regarding ClF3 as an alternative to BrF5 for
oxygen isotope analysis. Geochimica et Cosmochimica Acta 46, 1665-1668.
Bragg, W.L. & Claringbull, G.F. (1965): Crystal structures of minerals. Volume 4. Bell and
Sons Ltd., London, 149 pp.
Brandt, S., Klemd, R. & Okrusch, M. (2000): Evidence for ultra-high temperature
metamorphism in the Pre-Pan-African Epupa Complex, NW Namibia. GEOLUANDA
2000, Abstract volume: 33-34.
Brandt, S., Klemd, R. & Okrusch, M. (2003): Ultrahigh-temperature metamorphism and
multistage evolution of garnet-orthopyroxene granulites from the Proterozoic Epupa
Complex, NW Namibia. Journal of Petrology (in press).
328
21 References
Brandt, S., Will, T.M. & Klemd, R. (2001): Metamorphic evolution of sapphirine-bearing
orthopyroxene-sillimanite gneisses and associated felsic granulites from the Proterozoic
Epupa Complex, NW Namibia. EUG XI, Strasbourg, Abstract Volume: 571.
Brey, G.P. & Köhler, T (1990): Geothermobarometry in four-phase lherzolites II. New
thermobarometers, and practical assessment of existing thermobarometers. Journal of
Petrology 31: 1353-1378.
Brousse, R., Varet, J. & Bizouard, H. (1969): Iron in the minerals of the sodalite group.
Brown, P.E. (1989): FLINCOR - a microcomputer program for the reduction and investigation
of fluid inclusion data. American Mineralogist 74: 1390-1393.
Bühn, B. & Rankin, A.H. (1999): Composition of natural, volatile-rich Na-Ca-REE-Sr
carbonatitic fluids trapped in fluid inclusions. Geochimica et Cosmochimica Acta 63:
3781-3797.
Bühn, B., Wall, F. & Le Bas, M.J. (2001): Rare-earth element systematics of carbonatitic
fluorapatites, and their significance for carbonatite magma evolution. Contributions to
Cahen, L. & Snelling, N.J. (1966): The Geochronology of Equatorial Africa. Elsevier,
Amsterdam, 195 pp.
Carvalho, H. de & Alves, P. (1990): Gabbro-Anorthosite Complex of SW Angola/ NW
Namibia. Garcia de Orta, Série de Ciências da Terra, Communicações 2, 1-66.
Carvalho, H. de & Alves, P. (1993): The Precambrian of SW Angola and NW Namibia. Garcia
de Orta, Série de Ciências da Terra, Communicações 4: 1-38.
Carvalho, H. de, Crasto, J., Silva, Z.C.G. & Vialette, Y. (1987): The Kibaran cycle in Angola: a
discussion. Geological Journal 22: 85-102.
Carvalho, H. de, Fernandes, F.C. & Vialette, Y. (1979): Chronologie absolute du Précambrien
du Sud-Ouest de lÁngola. Comptes Rendus Académie Sciences Paris 288: 1647-1650.
Carvalho, H. de, Tassinari, C., Alves, P.H., Guimarães, F. & Simões, M.C. (2000):
Geochronological review of the Precambrian in western Angola: links with Brazil. Journal
of African Earth Sciences 31: 383-402.
Clifford, T.N. (1970): The structural framework of Africa. In: Clifford, T.N. & Gass, I.G. (eds.):
African magmatism and tectonics. Oliver and Boyd, Edinburgh, pp. 1-26.
Craig, J.R. (1971): Violarite stability relations. American Mineralogist 56: 1303-1311.
Craig, J.R. & Higgins, J.B. (1975): Cobalt- and iron-rich violarites from Virginia. American
Mineralogist 60: 35-38.
21 References
329
Craig, J.R. & Scott, S.D. (1974): Sulfide phase equilibria. In: Ribbe, P.H. (ed.): Sulfide
Mineralogy. Reviews in Mineralogy 1. Mineralogical Society of America, Washington
D.C., pp. CS1-CS110.
Craig, J.R., Vaughan, D.J. & Higgins, J.B. (1979): Phase relations in the Cu-Co-S system and
mineral associations of the carrollite (CuCo2S4) – linnaeite (Co3S4) series. Economic
Geology 74: 657-671.
Currie, K.L., Eby, G.L. & Gittins, J. (1986): The petrology of the Mont Saint Hilaire complex,
southern Quebec: An alkaline gabbro-peralkaline syenite association. Lithos 19: 65-81.
Currie, K.L. & Ferguson, J. (1971): A study of fenitization around the alkaline carbonatite
complex at Callander Bay, Ontario, Canada. Canadian Journal of Earth Sciences 8: 498518.
Dawson, J.B. (1989): Sodium carbonatite extrusions from Oldoinyo Lengai, Tanzania:
Implications for carbonatite complex genesis. In: Bell, K. (ed.): Carbonatites: genesis and
evolution. Unwin Hyman, London, pp. 255-277.
Dawson, J.B., Garson, M.S. & Roberts, B. (1987): Altered former alkalic carbonatite lava from
Oldoinyo Lengai, Tanzania: Inferences for calcite carbonatite lavas. Geology 15: 765-768.
Deer, W.A., Howie, R.A. & Zussman, J. (1963): Rock-forming minerals, Volume 4: Framework
silicates. Longmans, London, 435 pp.
Deer, W.A., Howie, R.A. & Zussman, J. (1992): An introduction to the rock-forming minerals,
2nd edition. Longmans, London, 696 pp.
Deines, P. (1989): Stable isotope variations in carbonatites. In: Bell, K. (ed.): Carbonatites:
genesis and evolution. Unwin Hyman, London, pp. 301-359.
de Waard, D. (1970): The anorthosite-charnockite suite of rocks of Roaring Brook Valley in the
eastern Adirondacks (Marcy Massif). American Mineralogist 55: 2063-2075.
Doherty, W. (1989): An internal standardization procedure for the determination of yttrium and
the rare earth elements in geological samples by inductively coupled plasma-mass
spectrometry. Spectrochimica Acta 44B: 263-280.
DÓrey, F.L.C. (1983): Cobaltopentlandite, siegenite ferifera e bravoite cobaltoniquelifera de
Mocambique. Garcia de Orta, Série de Geologia 6: 137-142.
Drüppel, K. (1999): Petrologie und Geochemie von Anorthositen des Kunene-IntrusivKomplexes, NW Namibia. Diploma thesis, University of Würzburg, 217 pp.
Drüppel, K., Hoefs, J. & Okrusch, M. (2002a): The polyphase magmatic and metasomatic
history of ferrocarbonatites from Swartbooisdrif, NW Namibia. Berichte der deutschen
Mineralogischen Gesellschaft, Beihefte zum European Journal of Mineralogy 13: 38.
330
21 References
Drüppel, K., Littmann, S. & Okrusch, M. (2000): Geo- und isotopenchemische Untersuchungen
der Anorthosite des Kunene-Intrusiv-Komplexes (KIK) in NW-Namibia. Berichte der
deutschen Mineralogischen Gesellschaft, Beihefte zum European Journal of Mineralogy
12: 37.
Drüppel, K. & Okrusch, M. (2000): Petrology and geochemistry of anorthositic rocks of the
Kunene Intrusive Complex, NW Namibia. GEOLUANDA 2000, Abstract volume: 49-50.
Drüppel, K., Rickers, K. & Okrusch, M. (2002b): Characterisation of the trace element
distribution in fenitizing fluids by synchrotron radiation x-ray fluorescence (SRXRF)
microprobe analysis. DESY Annual Report 2001 (online).
Drüppel, K., von Seckendorff, V. & Okrusch, M. (2001): Subsolidus reaction textures in
anorthosites of the Kunene Intrusive Complex, NW Namibia. European Journal of
Mineralogy 13: 289-309.
Duchesne, J.-C. (1984): Massif anorthosites: another partisan review. In: Brown W.L. (ed.):
Feldspars and feldspathoids. Reidel, Dordrecht, pp. 411-433.
Duchesne, J.-C. (1999): Fe-Ti deposits in Rogaland anorthosites (South Norway): geochemical
characteristics and problems of interpretation. Mineralium Deposita 34: 182-198.
Dulski, P. (1994): Interferences of oxide, hydroxide and chloride analyte species in the
determination of rare earth elements in geological samples by inductively coupled plasmamass spectrometry. Fresenius Journal of Analytical Chemistry 350:194-203.
Eckert, J.O. Jr., Newton, R.C. & Kleppa, O.J. (1991): The ∆H of reaction and recalibration of
garnet-pyroxene-plagioclase-quartz geobarometers in the CMAS system by solution
calorimetry. American Mineralogist 76: 148-160.
Emslie, R.F. (1975a): High pressure pyroxene megacrysts from anorthositic rocks and their
bearing on the genesis of the parent magmas. Geological Society of America, Abstracts
with Program 7: 752-753.
Emslie, R.F. (1975b): Major rock units of the Morin Complex, southwestern Quebec.
Geological Survey of Canada Papers 74-48: 1-37.
Emslie, R.F. (1978): Anorthosite massifs, rapakivi granites, and late Proterozoic rifting of North
America. Precambrian Research 7: 61-98.
Emslie, R.F. (1980): Geology and petrology of the Harp Lake Complex, Central Labrador: an
example of Elsonian magmatism. Geological Survey of Canada Bulletin 293:1-136.
Emslie, R.F. (1985): Proterozoic anorthosite massifs. In: Tobi, A. & Touret, J.L.R. (eds.): The
Deep Proterozoic Crust in the North Atlantic provinces, NATO ASI Series. Reidel,
Dordrecht, pp. 39-60.
21 References
331
Emslie, R.F. (1991): Granitoids of rapakivi granite-anorthosite and related associations.
Precambrian Research 51: 173-192.
Emslie, R.F., Hamilton, M.A. & Therault, R.J. (1994): Petrogenesis of mid-Proterozoic
anorthosite-mangerite-charnockite-granite, AMCG complex: isotopic and chemical
evidence from the Nain Plutonic Suite. Journal of Geology 102: 539-558.
Evenson, N.M., Hamilton, P.J. & OŃions, R.K. (1978): Rare-earth abundances in chondritic
meteorites. Geochimica et Cosmochimica Acta 42: 1199-1212.
Faure, G. (1986): Principles of isotope geology (2nd edition). John Wiley & Sons, New York,
589 pp.
Ferguson, J., McIver, J.R. & Danchlin, R.V. (1975): Fenitisation associated with the alkalinecarbonatite complex of Epembe, South West Africa. Transactions of the Geological
Society of South Africa 78: 111-121.
Ferry, J.M. & Spear, F.S. (1978): Experimental Calibration of the partitioning of Fe and Mg
between biotite and garnet. Contributions to Mineralogy and Petrology 66: 113-117.
Foster, M.D. (1960): Interpretation of the composition of trioctahedral micas. U.S. Geological
Survey, Professional Papers 254B: 24-49.
Franz, L. & Häussinger, H. (1990): Die Anwendung interaktiver Programme in der Mineralogie
am Beispiel des Programms “Amphibol”. Beiheft zum European Journal of Mineralogy 2:
68.
Freestone, I.C. & Hamilton, D.L. (1980): The role of liquid immiscibility in the genesis of
carbonatites – an experimental study. Contributions to Mineralogy and Petrology 73, 105117.
Frost, C.D., Frost, B.R., Bell, J.M. & Chamberlain, K.R. (2002): The relationship between Atype granites and residual magmas from anorthosite: evidence from the northern Sherman
batholith, Laramie Mountains, Wyoming, USA. Precambrian Research 119: 45-71.
Fuhrmann, M.L., Frost, B.R. & Lindsley, D.H. (1988): Crystallisation conditions of the Sybille
Monzosyenite, Laramie Anorthosite Complex, Wyoming. Journal of Petrology 29: 699729.
Fuhrmann, M.L. & Lindsley, D.H. (1988): Ternary feldspar modelling and thermometry.
American Mineralogist 73: 201-216.
Gaspar, J. & Wyllie, P.J. (1987): The phlogopites from the Jacupiranga carbonatite intrusions,
and their petrogenetic significance. Mineralogy and Petrology 36: 121-134.
Geiger, C.A., Newton R.C. & Klepper, O.J. (1987): Enthalpy of mixing of synthetic almandinegrossular and almandine-pyrope garnets from high-temperature solution calorimetry.
Geochimica et Cosmochimica Acta 51: 1755-1763.
332
21 References
Gittins, J. (1979): The feldspathoidal alkaline rocks. In: Yoder, H.S. Jr. (ed.): The Evolution of
the Igneous Rocks: Fiftieth Anniversary Perspectives. Princeton University Press,
Princeton, pp. 351-390.
Gittins, J. (1989): The origin and evolution of carbonatite magmas. In: Bell, K. (ed.):
Carbonatites: genesis and evolution. Unwin Hyman, London, pp. 580-600.
Gittins, J., Becket, M.F. & Jago, M.F. (1990): Composition of the fluid accompanying
carbonatite magma: a critical examination. American Mineralogist 75: 1106-1109.
Gittins, J. & Harmer, R.E. (1997): What is ferrocarbonatite? A revised classification. Journal of
African Earth Sciences 25, 159-168.
Grant, S.M. (1988): Diffusion models for corona formation in metagabbros from the Western
Grenville Province, Canada. Contributions to Mineralogy and Petrology 98: 49-63.
Grant, J.A. & Frost, B.R. (1990): Contact metamorphism and partial melting of pelitic rocks in
the aureole of the Laramie Anorthosite Complex, Morton Pass, Wyoming. American
Journal of Science 290: 425-472.
Green, D.H. & Ringwood, A.E. (1967): An experimental investigation of the gabbro to eclogite
transformation and its petrological applications. Geochimica et Cosmochimica Acta 31:
767-833.
Greig, J.W. & Barth, T.F.W. (1938): The system Na2O·Al2O3·2SiO2 (nepheline,carnegieite)Na2O·Al2O3·6SiO2 (albite). American Journal of Science 35A: 93-112.
Gresens, R.L. (1967): Composition-volume relationships of metasomatism. Chemical Geology
2: 47-65.
Griffin, W.L. (1971): Genesis of coronas in anorthosites of the Upper Jotun Nappe, Indre Sogn,
Norway. Journal of Petrology 12: 219-243.
Griffin, W.L. & Heier, K. (1973): Petrological implications of some corona structures. Lithos 6:
315-335.
Griffin, W.L., Mellini, M., Oberti, R. & Rossi, G. (1985): Evolution of coronas in Norwegian
anorthosites: re-evaluation based on crystal-chemistry and microstructures. Contributions
to Mineralogy and Petrology 91: 330-339.
Hamilton, D.L. (1964): Nephelines as crystallisation temperature indicators. Journal of Geology
9: 321-329.
Hammarstrom, J.M. & Zen, E.-A. (1983): Possible use of Al content in hornblende as a
geobarometer for plutonic rocks. Geological Society of America, Abstracts with Program
15: 590.
Hammarstrom, J.M. & Zen, E.-A. (1985): An empirical equation for igneous calcic amphibole
geobarometry. Geological Society of America, Abstracts with Program 17: 602.
21 References
333
Hammarstrom, J.M. & Zen, E.-A. (1986): Aluminium in hornblende: an empirical igneous
geobarometer. American Mineralogist 71: 1297-1313.
Harley, S.L. (1984): An experimental study of the partitioning of Fe and Mg between garnet and
orthopyroxene. Contributions to Mineralogy and Petrology 86: 359-373.
Hedberg, R.M. (1979): Stratigraphy of the Ovamboland Basin, South West Africa. Precambrian
Research Unit, University of Cape Town Bulletin 24: 1-325
Helz, R.T. (1973): Phase relations of basalts in their melting range at PH2O = 5 kbar as a function
of oxygen fugacity, I, Mafic phases. Journal of Petrology 14: 249-302.
Helz, R.T. (1976): Phase relations of basalts in their melting range at PH2O = 5 kbar, II, Melt
compositions. Journal of Petrology 17: 139-193.
Hey, M. (1954): A new review of chlorites. Mineralogical Magazine 30: 307-340.
Hibbard, M.J. (1994): Petrographic classification of crystal morphology. Journal of Geology
102: 571-581.
Hodges, K.V. & Spear, F.S. (1982): Geothermometry, geobarometry, and the Al2SiO5 triple
point at Mt. Moosilauke, New Hampshire. American Mineralogist 67: 1118-1134.
Hoefs J. (1997): Stable isotope geochemistry (2nd edition). Springer Verlag, Berlin, 208 pp.
Hoffman, P.F. (1989): Speculations on Laurentia`s first gigayear (2.0 to 1.0 Ga). Geology 17:
135-138.
Hogarth, D.D. (1977): Classification and nomenclature of the pyrochlore group. American
Hogarth, D.D. (1989): Pyrochlore, apatite and amphibole: distinctive minerals in carbonatite. In:
Bell, K. (ed.): Carbonatites: genesis and evolution. Unwin Hyman, London, pp. 105-148.
Holdaway, M.J. (1972): Thermal stability of Al-Fe epidote as a function of fo2 and Fe-content.
Holland, H.D. (1959): Some applications of thermochemical data to problems of ore deposits. I.
Stability relations among the oxides, sulphides, sulfates and carbonates of ore and gangue
metals. Economic Geology 54: 184-233
Hollister, L.S., Grissom, G.C., Peters, E.K., Stowell, H.H. & Sissom, V.B. (1987): Confirmation
of the empirical correlation of Al in hornblende with pressure of solidification of calcalkaline plutons. American Mineralogist 72: 231-239.
Horning-Kjarsgaard, I. (1998): Rare earth elements in sövitic carbonatites and their mineral
phases. Journal of Petrology 39: 2105-2122.
Horstmann, U.E. & Verwoerd, W.J. (1997): Carbon and oxygen isotope variations in southern
African carbonatites. Journal of African Earth Sciences 25: 115-138.
334
21 References
Indares, A. & Martignole, J. (1985): Biotite-garnet geothermometry in the granulite facies: the
influence of Ti and Al in biotite. American Mineralogist 70: 272-278.
Ixer, R.A. & Stanley, C.J. (1996): Siegenite-bearing assemblages found at the Great Orme mine,
Llandudno, North Wales. Mineralogical Magazine 60: 978-982.
Jamtveit, B., Dahlgren, S. & Austrheim, H. (1997): High-grade contact metamorphism of
calcareous rocks from the Oslo Rift, southern Norway. American Mineralogist 82: 12411254.
Johnson, C.D. & Carlson, W.D. (1990): The origin of olivine-plagioclase coronas in
metagabbros from the Adirondack Mountains, New York. Journal of metamorphic
Geology 8: 697-717.
Johnson, C.D. & Essene, E. (1982): The formation of garnet in olivine-bearing metagabbros
from the Adirondacks. Contributions to Mineralogy and Petrology 81: 240-251.
Keller, J. & Hoefs, J. (1995): Stable isotope characteristics of recent natrocarbonatites from
Oldoinyo Lengai. In: Bell, K. & Keller, J. (eds.): Carbonatite volcanism: Oldoinyo Lengai
and the petrogenesis of natrocarbonatites. Springer Verlag, Berlin, pp. 113-123.
Kerrick, D.M. & Jackobs, G.K. (1981): A modified Redlich-Kwong equation for H2O, CO2 and
H2O-CO2 mixtures at elevated temperatures and pressures. American Journal of Science
281: 735-767
Kjarsgaard, B.A. & Hamilton, D.L. (1989): The genesis of carbonatites by immiscibility. In:
Bell, K. (ed.): Carbonatites: genesis and evolution. Unwin Hyman, London, pp. 388-404.
Kleemann, U. & Reinhardt, J. (1994): Garnet-biotite thermometry revisited: The effect of AlVI
and Ti in biotite. European Journal of Mineralogy 6: 925-941.
Köstlin, E.C. (1967): The geology of part of the Kunene basic complex, Kaokoveld, South West
Africa. M.Sc. thesis, University of Cape Town, 92 pp.
Köstlin, E.C. (1974): The Kunene basic complex, northern South West Africa. In: Kröner, A.
(ed.): Contributions to the Precambrian geology of southern Africa. University of Cape
Town Bulletin 15, pp. 123-135.
Kolker, A. & Lindsley, D.H. (1989): Geochemical evolution of the Maloin Ranch pluton,
Laramie Anorthosite Complex, Wyoming: petrology and mixing relations. American
Kostelńikov, A.R. & Zhornyak, L.V. (1995): Stability of sodalite under hydrothermal
conditions. Geochemistry International 32: 87-90.
Kreher, B. (1992): Petrologie und Geochemie der Gabbrointrusion des Frankensteins
(Odenwald). Dr. rer. nat. thesis, University of Würzburg, 229 pp.
21 References
335
Kresten, P. (1988): The chemistry of fenitization: Examples from Fen, Norway. Chemical
Geology 68: 329-349.
Kresten, P. & Morogan, V. (1986): Fenititzation at the Fen complex, southern Norway. Lithos
19: 27-42.
Kretz, R. (1983): Symbols for rock-forming minerals. American Mineralogist 68: 277-279.
Kyser, T.K. (1986): Stable isotope variations in the mantle. In: Valley, J.W. (ed.): Stable isotopes
in high temperature geological processes. Reviews in Mineralogy 16. Mineralogical
Society of America, Washington D.C., pp. 141-164.
Laird, J. & Albee, A.L. (1981): High-pressure metamorphism in mafic schist from Northern
Vermont. American Journal of Science 281: 97-126.
Large, R.R., Gemmell, J.B., Paulick, H. & Huston, L. (2001): The alteration box plot: a simple
approach to understanding the relationship between alteration mineralogy and
lithogeochemistry associated with volcanic-hosted massive sulfide deposits. Economic
Geology 26: 957-971.
Leake, B.E., Schumacher, J.C., Smith, D.C., Stephenson, N.C.N., Ungaretti, L., Whittaker,
E.J.W. & Youzhi, G. (1997): Nomenclature of amphiboles. European Journal of
Mineralogy 9: 623-642.
Le Bas, M.J. (1977): Carbonatite-nephelinite volcanism: an African case history. Wiley,
London, 347 pp.
Le Bas, M.J. (1981): Carbonatite magmas. Mineralogical Magazine 44: 133-140.
Le Bas, M.J. (1987): Nephelinites and carbonatites. In: Fitton, J.G. & Upton, B.G.J. (eds.):
Alkaline igneous rocks. Geological Society Special Publications 30, Blackwell, Oxford, pp.
53-84.
Le Bas, M.J. (1989): Diversification of carbonatite. In: Bell, K. (ed.): Carbonatites: genesis and
Le Bas, M.J. (1999): Ferrocarbonatites: geochemistry and magma-fluid state. Memoirs of the
Geological Society of India 43, 785-802.
Le Bas, M.J., Keller, J., Tao, K., Wall, F, Williams, C.T. & Zhang, P. (1992): Carbonatite
dykes at Bayan Obo, Inner Mongolia, China. Mineralogy and Petrology 46: 195-228.
Lee, H.Y. & Ganguly, J. (1988): Equilibrium compositions of coexisting garnet and
orthopyroxene: Experimental determinations in the system FeO-MgO-Al2O3-SiO2, and
applications. Journal of Petrology 29: 93-113.
Leegaard, T. & Rosenqvist, T. (1964): Der Zersetzungsdruck und die Phasengleichgewichte der
höheren Sulfide von Kobalt und Nickel. Zeitschrift für Anorganische und Allgemeine
Chemie 328: 294-298.
336
21 References
Le Maitre, R.W. (ed.) (1989): A classification of igneous rocks and glossary of terms. Blackwell,
Oxford, 193 pp.
Lindsley, D.H. (1976): Experimental study of oxide minerals. In: Rumble, D. (ed.): Oxide
minerals, Reviews in Mineralogy 3. Mineralogical Society of America, Washington D.C.,
pp. L61-L88.
Lindsley, D.H. (1983): Pyroxene thermometry. American Mineralogist 68: 477-493.
Littmann, S., Cook, N.J., Teigler, B. & Drüppel, K. (2003): Ultramafic-mafic intrusives hosting
Cu-Ni sulphide mineralisation, Itjitambi area, Kunene Region, N.W. Nambia: Investigation
of drillcore, Hole OTD-1. Communications of the Geological Survey of Namibia (in press).
Littmann, S., Romer, R.L. & Okrusch, M. (2000): Nephelinsyenite der Epembe-SwartbooisdrifAlkali-Provinz (ESAP)/ NW Namibia. Berichte der deutschen Mineralogischen
Gesellschaft, Beihefte zum European Journal of Mineralogy 12: 115.
Longhi, J. & Ashwal, L.D. (1985): Two-stage models for lunar and terrestrial anorthosites:
Petrogenesis without a magma ocean. Journal of Geophysical Research 90: C571-C584.
Longhi, J., Vander Auwera, J., Fram, M.S. & Duchesne, J.-C. (1999): Some Phase
Equilibrium Constraints on the Origin of Proterozoic (Massif) Anorthosites and Related
Rocks. Journal of Petrology 40: 339-362.
Ludington, S. (1978): The biotite-apatite geothermometer revisited. American Mineralogist 63:
551-553.
Markl, G., Marks, M., Schwinn, G. & Sommer, H. (2001): Phase equilibrium constraints on
intensive crystallization parameters of the Ilímaussaq complex, South Greenland. Journal
of Petrology 42: 2231-2258.
Martignole, J. (1974): L`évolution magmatique du complexe de Morin et son apport au
problème des anorthosites. Contributions to Mineralogy and Petrology 44: 117-137.
Martignole, J. & Schrijver, K. (1971): Association of (hornblende-) garnet-pyroxene
”subfacies” of metamorphism and anorthosite mass. Canadian Journal of Earth Sciences
8: 698-704.
Martin, H. (1965): The Precambrian geology of South West Africa and Namaqualand.
Precambrian Research Unit, University of Cape Town, 159 pp.
Mayer, A., Sinigoi, S., Miguel, L.G., Morais, E. & Petrini, R. (2000): Kibaran ages in the
Kunene anorthositic complex. GEOLUANDA 2000, Abstract volume: 106.
McLelland, J.M. & Whitney, P.R. (1977): The origin of garnet in the anorthosite-charnockite
suite of the Adirondacks. Contributions to Mineralogy and Petrology 60: 61-181.
21 References
337
McLelland, J.M. & Whitney, P.R. (1980a): A geological garnet-forming reaction for metaigneous rocks in the Adirondacks. Contributions to Mineralogy and Petrology 72: 111122.
McLelland, J.M. & Whitney, P.R. (1980b): Compositional controls on spinel clouding and
garnet formation in plagioclase of olivine metagabbros, Adirondack Mountains, New
York. Contributions to Mineralogy and Petrology 73: 243-251.
Menge, G.F.W. (1986): Sodalite-carbonatite deposits of Swartbooisdrif, South West
Africa/Namibia. In: Anhaeusser, C.R. & Maske, S. (eds.): Mineral Deposits of Southern
Africa. Geological Society of South Africa, Johannesburg, pp. 2261-2268.
Menge, G.F.W. (1996): The eastern portion of the Kunene complex, its satellite intrusions and
the alkaline suite between Epembe and Swartbooisdrif. Unpublished Report, 128 pp.
Menge, G.F.W. (1998): The antiformal structure and general aspects of the Kunene Complex.
Zeitschrift der Deutschen Geologischen Gesellschaft 149: 431-448.
Minceva-Stefanova, J. (1978): Paragenetische und geochemische Bedeutung der Mineralien von
den Bravoit-, Cobaltin- und Linneit-Gruppen in den Polymetallvererzungen des "StrataBound"-Typs in den Balkaniden. Verhandlungen der Geologischen Bundesanstalt, Wien:
419-435.
Minceva-Stefanova, J. & Kostov, I. (1976): On siegenite and the `miscibility´ between linnaeite
and polydymite. Geochemistry, Mineralogy and Petrology, Sofia 4: 35-56.
Moecher, D.P., Essene, E.J. & Anovitz, L.M. (1988): Calculation and application of
clinopyroxene-garnet-plagioclase-quartz geobarometers. Contributions to Mineralogy and
Mora, C.I. & Valley, J.W. (1988): Channelized meteoric-hydrothermal alteration of the Boehls
Butte and Bitterroot anorthosites, northern ID and MT. Eos 69: 1489.
Morais, E., Sinigoi, S., Mayer, A., Mucana, A., Miguel, L.G. & Rufino Neto, J. (1998): The
Kunene gabbro-anorthosite Complex: preliminary results based on new field and chemical
data. African Geoscience Review 5: 485-498.
Morimoto, N. (1988): Nomenclature of pyroxenes. Mineralogical Magazine 52: 535-550.
Morogan, V. (1982): Fenitization and ultimate rheomorphism of xenoliths from the Oldoinyo
Lengai carbonatitic volcano, Tanzania. M.Sc. thesis, University of Montreal, 193 pp.
Morogan, V. (1989): Mass transfer and mobility of REE during fenitization at Alnö, Sweden.
Morogan, V. (1994): Ijolite vs. Carbonatite as sources of fenitization. Terra Nova 6: 166-176.
338
21 References
Morogan, V. & Lindblom, S. (1995): Volatiles associated with the alkaline-carbonatite
magmatism at Alnö, Sweden: A study of fluid and solid inclusions in minerals from the
Langarsholmen ring complex. Contributions to Mineralogy and Petrology 122: 262-274.
Morogan, V. & Martin, R.F. (1985): Mineralogy and partial melting of fenitized crustal
xenoliths in the Oldoinyo Lengai carbonatitic volcano, Tanzania. American Mineralogist
70: 1114-1126.
Morogan, V. & Woolley, A.R. (1988): Fenitization in the Alnö complex, Sweden: distribution,
mineralogy and genesis. Contributions to Mineralogy and Petrology 100: 169-182.
Morrison, J. & Valley, J.W. (1988): Contamination of the Marcy anorthosite massif,
Adirondack Mountains, NY: petrologic and isotopic evidence. Contributions to Mineralogy
and Petrology 98: 97-108.
Morse, S.A. (1968): Layered intrusions and anorthosite genesis. In: Isachsen, Y.W. (ed.): Origin
of anorthosite and related rocks. N.Y. State Museum and Science Service, Memoir 18, pp.
175-187.
Morse, S.A. (1982): A partisan review of Proterozoic anorthosites. American Mineralogist 67:
1087-1100.
Morse, S.A. & Nolan, K.M. (1984): Origin of strongly reversed rims on plagioclase in
cumulates. Earth and Planetary Science Letters 68: 485-498.
Munoz, J.L. & Swenson, A. (1981): Chloride-hydroxyl exchange in biotite and estimation of
relative HCl/HF activities in hydrothermal fluids. Economic Geology 76: 2212-2221.
Nelson, D.R., Chivas, A.R., Chappell, B.W. & Mc Culloch, M.T. (1988): Geochemical and
isotope systematics in carbonatites and implications for the evolution of ocean-island
sources. Geochimica et Cosmochimica Acta 52: 1-17.
Nesbitt, B.E. & Kelley, W.C. (1977): Magmatic and hydrothermal inclusions in carbonatite of
the Magnet Cove complex, Arkansas. Contributions to Mineralogy and Petrology 63: 271294.
Newton, R.C. & Perkins, D. (1982): Thermodynamic calibration of geobarometers based on the
assemblages garnet–plagioclase–orthopyroxene(clinopyroxene)-quartz. American Mineralogist 67: 203-222.
Nickel, H.E. (1970): The application of ligand field concepts to an understanding of the
structural stabilities and solid solution limit of sulfides and related minerals. Chemical
Geology 5: 233-241.
Oberti, R., Smith, D.C., Rossi, G. & Caucia, F. (1991): The crystal-chemistry of highaluminium titanites. European Journal of Mineralogy 3: 777-792.
21 References
339
OĆonnor, Y.-L. & Morrison, J. (1999): Oxygen isotope constraints on the petrogenesis of the
Sybille intrusion of the Proterozoic Laramie anorthosite complex. Contributions to
Ohmoto, H. & Goldhaber, M.B. (1997): Sulphur and carbon isotopes. In: Barnes, H.L. (ed.):
Geochemistry of hydrothermal ore deposits. John Wiley & Sons, New York, pp. 517-611.
Olson, K.E. & Morse, S.A. (1990): Regional Al-Fe mafic magmas associated with anorthositebearing terranes. Nature 344: 760-762.
Ostwald, J. (1978): Linnaeite series minerals from the Kalgoorlie district, western Australia.
Mineralogical Magazine 42: 93-98.
Otten, M.T. (1984): The origin of brown hornblende in the Artfjället gabbro and dolerites.
Papike, J.J. & Cameron, M. (1980): Crystal chemistry of silicate pyroxenes. In: Prewitt, C.T.
(ed.): Pyroxenes. Reviews in Mineralogy 7. Mineralogical Society of America, Washington
D.C., pp. 5-92.
Papike, J.J., Cameron, K.L. & Baldwin, K. (1974): Amphiboles and pyroxenes:
Characterization of other than quadrilateral components and estimates of ferric iron from
microprobe data. Geological Society of America, Abstracts with Program 6: 1053-1054.
Pearce, J.A., Harris, N.B.W. & Tindle, A.G. (1984): Trace element discrimination diagrams for
the tectonic interpretation of granitic rocks. Journal of Petrology 25: 956-983.
Pearce, N.J.G., Perkins, W.T., Westgate, J.A., Gorton, M.P., Jackson, S.E., Neal, C.R. &
Chenery, S.P. (1997): A compilation of new and published major and trace element data
for NIST SRM 610 and NIST SRM 612 glass reference materials. Geostandards
Newsletter 21: 115-144.
Peck, W.H. & Valley, J.W. (2000): Large crustal input into high δ18O anorthosite massifs of the
southern Grenville Province: new evidence from the Morin Complex, Quebec.
Perchuk, L.L. & Lavrentéva, I.V. (1983): Experimental investigation of exchange equilibria in
the system cordierite-garnet-biotite. In: Saxena, S.K. (ed.): Kinetics and equilibrium in
mineral reactions. Advances in Physical Geochemistry, 3. Springer Verlag, New York, pp.
199-239.
Pignolet, S. & Hagni, R.D. (1983): Cobalt-nickel mineralization associated with lead-zinccopper mineralization in the Mississippi Valley-type deposits at Fredericktown, Missouri.
In: Kisvarsanyi, G., Grant, S.K., Pratt, W.P. & Koenig, J.W. (eds.): International
conference on Mississippi Valley Type lead-zinc deposits. University of Missouri-Rolla,
Missouri-Rolla, pp.187-194.
340
21 References
Poldervaart, A. & Hess, H.H. (1951): Pyroxenes in the crystallisation of basaltic magmas.
Journal of Geology 59: 472-489.
Powell, R. & Holland, T.J.B. (1988): An internally consistent dataset with uncertainties and
correlations. 3. Applications to geobarometry, worked examples and a computer program.
Journal of metamorphic Geology 6: 173-204.
Rae, D.A., Coulson, I.M. & Chambers, A.D. (1996): Metasomatism in the North Qôroq centre,
South Greenland: apatite chemistry and rare-earth element transport. Mineralogical
Magazine 60: 207-220.
Rankin, A.H. (1975): Fluid inclusion studies in apatite from carbonatites of the Wasaki area of
western Kenya. Lithos 8: 123-136.
Riley, J.F. (1980): Ferroan carrollites, cobaltian violarites, and other members of the linnaeite
group: (Co,Ni,Fe,Cu)3S4. Mineralogical Magazine 43: 733-739.
Roedder, E. (1973): Fluid inclusions from the fluorite deposits associated with carbonatite of
Amba Dongar, India and Okoruso, South West Africa. Transactions of the Institution of
Mining and Metallurgy B 82: 835-839.
Roedder, E. (1984): Fluid inclusions. Reviews in Mineralogy 12. Mineralogical Society of
America, Washington D.C.,646 pp.
Rønsbo, J.G. (1989): Coupled substitution involving REEs and Na and Si in apatites in alkaline
rocks from the Ilímaussaq intrusion, South Greenland, and its petrological implications.
Rosenqvist, T. (1954): A thermodynamic study of the iron, cobalt, and nickel sulphides. Journal
of the Iron and Steel Institute 176: 37-57.
Samson, I.M., Weining, L. & Williams-Jones, A.E. (1995a): The nature of orthomagmatic
hydrothermal fluids in the Oka carbonatite, Quebec: evidence from fluid inclusions.
Samson, I.M., Williams-Jones, A.E. & Weining, L. (1995b): The chemistry of hydrothermal
fluids in carbonatites: evidence from leachate and SEM-decrepitate analysis of fluid
inclusions from Oka, Quebec, Canada. Geochimica et Cosmochimica Acta 59: 1979-1989.
Schliestedt, M. (1980): Phasengleichgewichte in Hochdruckgesteinen von Sifnos, Griechenland.
PhD thesis, University of Braunschweig, 142 pp.
Seifert, K.E. (1978): Anorthosite-mangerite relations on Barker Mountain, New York.
Geological Society of America, Bulletin 89: 245-250.
Seifert, W., Kämpf, H. & Wasternack, J. (2000): Compositional variation in apatite, phlogopite
and other accessory minerals of the ultramafic Delitzsch complex, Germany: implication
for cooling history of carbonatites. Lithos 53: 81-100.
21 References
341
Seim, R. & Tischendorf, G. (1990): Grundlagen der Geochemie. Deutscher Verlag für
Grundstoffindustrie, Leipzig, 632 pp.
Seth, B., Brandt, S. & Kramers, J.B. (2001): First isotopic age determinations of midProterozoic granulite-facies metamorphism in the Epupa Complex. EUG XI, Strasbourg,
Abstract Volume: 597.
Sharp, Z.D., Helfrich, G.R., Bohlen, S.R. & Essene, E.J. (1989): The stability of sodalite in the
system NaAlSiO4-NaCl. Geochimica et Cosmochimica Acta 53: 1943-1954.
Shepherd, T. J., Rankin, A. H. & Alderton, D. H. M. (1985): A Practical Guide to Fluid
Inclusion Studies. Blackie, Glasgow & London, 239 pp.
Silva, Z.C.G. (1990): Geochemistry of the gabbro-anorthosite Complex of Angola. Journal of
African Earth Sciences 10: 683-692.
Silva, Z.C.G. (1992). Mineralogy and cryptic layering of the Kunene anorthosite complex of SW
Angola and Namibia. Mineralogical Magazine 56: 319-327.
Simpson, E.S.W. (1970): The anorthosite of southern Angola: a review of present data. In:
Clifford, T.N. & Gass, I.G. (eds.): African magmatism and tectonics. Oliver Boyd,
Edinburgh, pp. 89-96.
Simpson, E.S.W. & Otto, J.D.T. (1960): On the Pre-Cambrian anorthosite mass of southern
Angola. Proceedings of the 21st IGC, Part 13 (Petrographic provinces, igneous and
metamorphic rocks), Copenhagen: 216-227.
Slejko, F., Demarchi, G. & Morais, E. (2002): Mineral chemistry and Nd isotopic composition
of two anorthositic rocks from the Kunene complex (South Western Angola). Journal of
African Earth Sciences 35: 77-88.
Sörensen, H. (1970): Internal structures and geological setting of the three agpaitic intrusions Khibina and Lovozero of the Kola Peninsula and Ilímaussaq, south Greenland. Canadian
Spear, F.S. & Kimball, K.L. (1984): RECAMP-A FORTRAN IV program for estimating
Fe3+contents in amphiboles. Computers in Geology 10: 317-325.
Spencer, K.J. & Lindsley, D.H. (1978): New experimental results on the magnetite-ulvöspinel,
hematite-ilmenite solution model using the Co-CoO buffer. Geological Society of America,
Abstracts with Program 10: 496.
Spencer, K.J. & Lindsley, D.H. (1981): A solution model for coexisting iron-titanium oxides.
Stone, P. & Brown, G.M. (1958): The Quihita-Cunene layered gabbroic intrusion of South-West
Angola. Geological Magazine 95: 195-206.
342
21 References
Stormer, J.C. & Carmichael, I.S.E. (1971): Fluorine-hydroxyl exchange in apatite and biotite: a
potential igneous geothermometer. Contributions to Mineralogy and Petrology 31, 121131.
Taylor, H.P. (1968): Oxygen isotope studies of anorthosites, with particular reference to the
origin of bodies in the Adirondack Mountains, New York. In: Isachsen, Y.W. (ed.): Origin
of anorthosite and related rocks. N.Y. State Museum and Science Service, Memoir 18, pp.
111-134.
Taylor, H.P., Frechen, J. & Degens, E.T. (1967): Oxygen and carbon isotope studies of
carbonatites from the Laacher See district, West Germany and the Alnö district, Sweden.
Taylor, H.P. & Sheppard, S.M.F. (1986): Igneous rocks I. Processes of isotopic fractionation
and isotope systematics. In: Valley, J.W., Taylor, H.P. & OŃeil, J. (eds.): Stable isotopes
in high temperature geological processes. Mineralogical Society of America, Chelsea,
Michigan, pp. 227-269.
Taylor, S.R., Campbell, I.H., McCulloch, M.T. & McLennan S.M. (1984): A lower crustal
origin for massif-type anorthosites. Nature 311: 372-374.
Tegtmeyer, A. & Kröner, A. (1985): U-Pb zircon ages for granitoid gneisses in northern Namibia
and their significance for Proterozoic crustal evolution of Southwestern Africa.
Precambrian Research 28: 311-326.
Thompson, R.N. (1982): Magmatism of the British Tertiary Volcanic Province. Scottish Journal
of Geology 18:49-107.
Thompson, R.N., Morrison, M.A., Dickin, A.P. & Hendry, G.L. (1983): Continental flood
basalts … arachnids rule OK?. In: Hawkesworth, C.J. & Norry, M.J. (eds.): Continental
basalts and mantle xenoliths. Shiva Publishing Ltd, Nantwich, pp. 158-185.
Thompson, R.N., Smith, P.M., Gibson, S.A., Mattey, D.P. & Dickin, A.P. (2002): Ankerite
carbonatite
from
Swartbooisdrif,
Namibia:
the
first
evidence
for
magmatic
ferrocarbonatite. Contributions to Mineralogy and Petrology 143: 377-396.
Torquato, J.R. & Carvalho, J.A.R. (1992): Idade Rb-Sr do granito do Carculo, uma nova
evidência para a existência do evento Namib no Sudoeste de Angola. Revista Geologia 50:
157-167.
Torquato, J.R. & Oliviera, J.T. (1977): Sobre a idade dos granitos e do grupo vulcanosedimentar da região de Chipindo, Angola. Communicações Serviços Geológicos Portugal
61: 223-238.
21 References
343
Torquato, J.R., Silva, A.T.S.F. da, Cordani, U.G. & Kawashita, K. (1979): A evolução
geológica do Cinturão Móvel do Quipungo no Ocidente de Angola. Anais Academia
Brasileira Ciências 51: 133-143.
Toulmin, P. III & Barton, P.B. Jr. (1964): A thermodynamic study of pyrite and pyrrhotite.
Treiman, A.H. & Essene, E.J. (1985): The Oka carbonatite complex, Quebec: geology and
evidence for silicate-carbonate liquid immiscibility. American Mineralogist 70: 1101-1113.
Tröger, W.E. (1982): Optische Bestimmung der gesteinsbildenden Minerale, Teil 2.
Schweizerbartsche Verlagsbuchhandlung, Stuttgart, 822 pp.
Trompette, R. (1994): Geology of Western Gondwana (2000-500 Ma). Pan-African-Brasiliano
aggregation of South America and Africa. Balkema, Rotterdam, 350 pp.
Tuttle, O.F. & Smith, J.V. (1958): The nepheline-kalsilite system: II. Phase relations. American
Journal of Science 256: 571589.
Valley, J.W. & OŃeil, J.R. (1982): Oxygen isotope evidence for shallow emplacement of
Adirondack anorthosite. Nature 300: 497-500.
Vaughan, D.J., Burns, R.G. & Burns, V.M. (1971): Geochemistry and bonding of thiospinel
minerals. Geochimica et Cosmochimica Acta 35: 365-381.
Vaughan, D.J. & Craig, J.R. (1985): The crystal chemistry of iron-nickel thiospinels. American
Vermaak, C.F. (1981): Kunene Anorthosite Complex. In: Hunter, D.R. (ed.): Precambrian of
the Southern Hemisphere. Developments in Precambrian Geology, Vol. 2. Elsevier,
Amsterdam, pp. 578-599.
Vertainen, H. & Woolley, A.R. (1974): The age of the Sokli carbonatite, Finland, and some
relationships of the North Atlantic alkaline igneous province. Bulletin of the Geological
Society of Finland 46: 81-91.
Verwoerd W.J. (1967): The carbonatites of South Africa and South West Africa. Handbook of
the Geological Survey of South Africa 6: 1-452.
Verwoerd, A.J. (1993): Update on carbonatites of South Africa and Namibia. South African
Journal of Geology 96: 75-95.
Volfinger, M., Robert, J.-L., Vielzeuf, D. & Neiva, A.M.R. (1985): Structural control of the
chlorine content of OH-bearing silicates (micas and amphiboles). Geochimica et
Von Seckendorff, V. & Drüppel, K. (1999): Fe-Ti oxides and sulphide mineralisation in
sodalite-bearing metasomatites at Swartbooisdrif (Kunene Region, NW Namibia). In: de
344
21 References
Wall, H. & Greiling, R.O. (eds.): Aspects of Pan-African tectonics. Bilateral Seminars 32.
Forschungszentrum Jülich GmbH, Jülich, pp. 99-102.
Von Seckendorff, V., Drüppel, K., Okrusch, M., Cook, N. & Littmann, S. (2000): Oxidesulphide relationships in sodalite-bearing metasomatites of the Epembe-Swartbooisdrif
alkaline province, north-west Namibia. Mineralium Deposita 35, 430-450.
Wagner,
T.
(1999):
Spätvaristische
hydrothermale Mineralisationen im Rheinischen
Schiefergebirge. Freiberger Forschungshefte, Freiberg C 478: VIII + 194 pp.
Wall, F., Le Bas, M.J. & Srivastava, R.K. (1993): Calcite and carbocernaite exsolution and
cotectic textures in a Sr, REE-rich carbonatite dyke from Rajasthan, India. Mineralogical
Magazine 57: 495-513.
Watkinson, D.H. & Wyllie, P.J. (1971): Experimental study of the join NaAlSiO4-CaCO3-H2O
and the genesis of alkalic rock-carbonatite complexes. Journal of Petrology 12: 357-378.
Wellman, T.R. (1970): The stability of sodalite in a synthetic syenite plus aqueous chloride fluid
system. Journal of Petrology 11: 49-71.
Wen, S. & Nekvasil, H (1994): SOLVCALC: an interactive graphics program package for
calculating the ternary feldspar solvus and for two-feldspar geothermometry. Computers &
Geosciences 20: 1025-1040.
Whalen, J.B., Currie, K.L. & Chappell, B.W. (1987): A-type granites: geochemical
characteristics, discrimination and petrogenesis. Contributions to Mineralogy and
Whitney, P.R. & McLelland, J.M. (1973): Origin of coronas in metagabbros of the Adirondack
Mts., N.Y. Contributions to Mineralogy and Petrology 39: 81-98.
Wiebe, R.A. (1986): Lower crustal cumulate nodules in Proterozoic dykes of the Nain complex,
Labrador: evidence for the origin of Proterozoic anorthosites. Journal of Petrology 27:
1253-1275.
Wiebe, R.A. (1994): Proterozoic anorthosite complexes. In: Condie, K.C. (ed.): Proterozoic
crustal Evolution. Elsevier, Amsterdam, pp. 215-261.
Williams-Jones, A.E. & Palmer, D.A.S. (2002): Fluid evolution of the Amba Dongar carbonatite
complex, India. Chemical Geology 185: 283-301.
Woolley, A.R. (1989): The spatial and temporal distribution of carbonatites. In: Bell, K. (ed.):
Carbonatites: genesis and evolution. Unwin Hyman, London, pp. 15-37.
Woolley, A.R. & Kempe, D.R.C. (1989): Carbonatites: nomenclature, average chemical
compositions and element distribution. In: Bell, K. (ed.): Carbonatites: genesis and
21 References
345
Wyllie, P.J. (1966): Experimental studies of carbonatite problems: The origin and differentiation
of carbonatite magmas. In: Tuttle, O.F. & Gittins, J. (eds.): Carbonatites. Interscience,
New York, pp. 311-352.
Wyllie, P.J. (1989): Origin of carbonatites: evidence from phase equilibrium studies. In: Bell, K.
(ed.): Carbonatites: genesis and evolution. Unwin Hyman, London, pp. 149-176.
Wyllie, P.J. & Tuttle, O.F. (1960): The system CaO-CO2-H2O and the origin of carbonatites.
Journal of Petrology 1: 1-46.
Yang, H.-Y. (1987): Stability of ilmenite and titanomagnetite in the presence of carbon dioxide a
thermodynamic evaluation. Contributions to Mineralogy and Petrology 95: 202-206.
Yoder, H.S. & Kushiro, I. (1969): Melting of a hydrous phase: phlogopite. American Journal of
Science 267-A, 558-582.
Yund, R.A. & Kullerud, G. (1966): Thermal stability of assemblages in the Cu-Fe-S system.
Journal of Petrology 7: 454-488
Zhang, Y. & Franz, J.D. (1987): Determination of the homogenization temperatures and
densities of supercritical fluids in the system NaCl-KCl-CaCl2-H2O using synthetic fluid
inclusions. Chemical Geology 64: 335-350.
Zheng, Y.F. (1991): Calculation of oxygen isotope fractionation in metal oxides. Geochimica et
Zheng, Y.F. (1999): Oxygen isotope fractionation in carbonate and sulfate minerals
Geochemical Journal 33: 109-126.
Zhu, C. & Sverjensky, D.A. (1991): Partitioning of F-Cl-OH between minerals and hydrothermal
fluids. Geochimica et Cosmochimica Acta 55: 1837-1858.
Zhu, C. & Sverjensky, D.A. (1992): F-Cl-OH partitioning between biotite and apatite.
Zuleger, E. & Erzinger, J. (1988): Determination of the REE and Y in silicate materials with
ICP-AES. Fresenius Zeitschrift für Analytische Chemie 332, 140-143.
Lebenslauf
Kirsten Drüppel
Virchowstr. 18
97072 Würzburg
Geburtsdatum:
01.06.1974
Geburtsort:
Würzburg
Familienstand:
ledig
Schulischer Werdegang
09/1984-06/1993
Siebold-Gymnasium Würzburg
06/1993
Abitur
Studium
10/1993-09/2000
Studium
der
Geologie-Paläontologie
an
der
Bayerischen
Julius-
Maximilians-Universität Würzburg
05/1997-10/1999
Studium der Mineralogie an der Bayerischen Julius-MaximiliansUniversität Würzburg
05/1996
Diplom-Vorprüfung im Studiengang Geologie-Paläontologie
11/1997
Diplom-Vorprüfung im Studiengang Mineralogie
01/1999
Diplomhauptprüfung im Studiengang Mineralogie
01/1999-10/1999
Diplomarbeit im Studiengang Mineralogie über das Thema “Petrologie
und Geochemie von Anorthositen des Kunene-Intrusiv-Komplexes, NWNamibia”
10/1999
Diplom
seit 10/2000
Promotionsstudium am Mineralogischen Institut der Universität Würzburg
Tätigkeiten als wissenschaftliche Hilfskraft
02/1997-06/1997
Wissenschaftliche Hilfskraft am geologischen Institut der Universität
Würzburg; Fachbereich: Strukturgeologie
08/1998-12/1998
Wissenschaftliche Hilfskraft am Mineralogischen Institut der Universität
Würzburg; Fachbereich: Archäometrie
Tätigkeiten als wissenschaftliche Angestellte
10/1999-03/2000
Wissenschaftliche Angestellte am Fraunhofer-Institut für Silicatforschung
(ISC), Würzburg; Fachbereich: Glas und Keramik
07/2000-03/2003
Wissenschaftliche Angestellte am Mineralogischen Institut der Universität
Würzburg; Fachbereich: metamorphe und magmatische Petrologie
Tätigkeiten als Lehrbeauftragte
09/2001-07/2003
Lehrbeauftragte am Institut für Geologie und Mineralogie der FriedrichAlexander-Universität, Erlangen; Lehrveranstaltungen: „Mikroskopie II:
gesteinsbildende Minerale“ (WS 2001/2001, SS 2002, WS 2002/2003, SS
2003) und „Mikroskopisches Praktikum IV: Auflichtmikroskopie“ (SS
2002)
Auslandsaufenthalte
03/1998-05/1998, 03/1999-05/1999 & 08/2001
Forschungsaufenthalte in Namibia

zusammenfassung - OPUS Würzburg

Transcription

Similar documents

How to Tie a Woggle

lnstallation - ConservationMart.com

Turkey - EOPOWER

Key - Scioly.org

RÉSoNANSS: a regional contribution to

Real Time Control of Hand Prosthesis Using Surface EMG

Kerntech Diagnose-System Loose Part Monitoring System

CoUNTerMEASURES CoUNTerMEASURES

PrimeFilm 1800 AFL Low-Cost 35mm Slide/Film Scanning Through

IC-910 SDR Mod